### Refine

#### Year of publication

- 1999 (425) (remove)

#### Document Type

- Preprint (351)
- Article (66)
- Report (4)
- Diploma Thesis (1)
- Lecture (1)
- Periodical Part (1)
- Study Thesis (1)

#### Language

- English (425) (remove)

#### Keywords

- AG-RESY (6)
- Case-Based Reasoning (6)
- HANDFLEX (5)
- Location Theory (5)
- PARO (5)
- case-based problem solving (5)
- Abstraction (4)
- Knowledge Acquisition (4)
- resolution (4)
- Fallbasiertes Schließen (3)

#### Faculty / Organisational entity

We present an approach to learning cooperative behavior of agents. Our ap-proach is based on classifying situations with the help of the nearest-neighborrule. In this context, learning amounts to evolving a set of good prototypical sit-uations. With each prototypical situation an action is associated that should beexecuted in that situation. A set of prototypical situation/action pairs togetherwith the nearest-neighbor rule represent the behavior of an agent.We demonstrate the utility of our approach in the light of variants of thewell-known pursuit game. To this end, we present a classification of variantsof the pursuit game, and we report on the results of our approach obtained forvariants regarding several aspects of the classification. A first implementationof our approach that utilizes a genetic algorithm to conduct the search for a setof suitable prototypical situation/action pairs was able to handle many differentvariants.

The common wisdom that goal orderings can be used to improve planning performance is nearly as old as planning itself. During the last decades of research several approaches emerged that computed goal orderings for different planning paradigms, mostly in the area of state-space planning. For partial-order, plan-space planners goal orderings have not been investigated in much detail. Mechanisms developed for statespace planning are not directly applicable because partial-order planners do not have a current (world) state. Further, it is not completely clear how plan-space planners should make use of goal orderings. This paper describes an approach to extract goal orderings to be used by the plan-space planner CAPlan. The extraction of goal orderings is based on the analysis of an extended version of operator graphs which previously have been found useful for the analysis of interactions and recursion of plan-space planners.

Freivalds, Karpinski and Smith [8] explored a special type of learning in the limit: identification of an unknown concept (function) by eliminating (erasing) all but one possible hypothesis (this type of learning is called co-learning). The motivation behind learning by erasing lies in the process of human and automated computer learning: often we can discard incorrect solutions much easier than to come up with the correct one. In Gödel numberings any learnable family can be learned by an erasing strategy. In this paper we concentrate on co-learning minimal programs. We show that co-learning of minimal programs, as originally defined is significantly weaker than learning minimal programs in Gödel numberings. In order to enhance the learning power

We present an approach to automating the selection of search-guiding heuris-tics that control the search conducted by a problem solver. The approach centerson representing problems with feature vectors that are vectors of numerical val-ues. Thus, similarity between problems can be determined by using a distancemeasure on feature vectors. Given a database of problems, each problem beingassociated with the heuristic that was used to solve it, heuristics to be employedto solve a novel problem are suggested in correspondence with the similaritybetween the novel problem and problems of the database.Our approach is strongly connected with instance-based learning and nearest-neighbor classification and therefore possesses incremental learning capabilities.In experimental studies it has proven to be a viable tool for achieving the finaland crucial missing piece of automation of problem solving - namely selecting anappropriate search-guiding heuristic - in a flexible way.This work was supported by the Deutsche Forschungsgemeinschaft (DFG).

This report presents the properties of a specification of the domain of process planning for rotary symmetrical workpieces. The specification results from a model for problem solving in this domain that involves different reasoners, one of which is an AI planner that achieves goals corresponding to machining workpieces by considering certain operational restrictions of the domain. When planning with SNLP (McAllester and Rosenblitt, 1991), we will show that the resulting plans have the property of minimizing the use of certain key operations. Further, we will show that, for elastic protected plans (Kambhampati et al., 1996) such as the ones produced by SNLP, the goals corresponding to machining parts of a workpiece are OE-constrained trivial serializable, a special form of trivial serializability (Barrett and Weld, 1994). However, we will show that planning with SNLP in this domain can be very difficult: elastic protected plans for machining parts of a workpiece are nonmergeable. Finally, we will show that, for sufix, prefix or sufix and prefix plans such as the ones produced by state-space planners, it is not possible to have both properties, being OEconstrained trivial serializable and minimizing the use of the key operations, at the same time.

The feature interaction problem in telecommunications systems increasingly obstructsthe evolution of such systems. We develop formal detection criteria which render anecessary (but less than sufficient) condition for feature interactions. It can be checkedmechanically and points out all potentially critical spots. These have to be analyzedmanually. The resulting resolution decisions are incorporated formally. Some prototypetool support is already available. A prerequisite for formal criteria is a formal definitionof the problem. Since the notions of feature and feature interaction are often used in arather fuzzy way, we attempt a formal definition first and discuss which aspects can beincluded in a formalization (and therefore in a detection method). This paper describeson-going work.

Collecting Experience on the Systematic Development of CBR Applications using the INRECA Methodology
(1999)

This paper presents an overview of the INRECA methodology for building and maintaining CBR applications. This methodology supports the collection and reuse of experience on the systematic development of CBR applications. It is based on the experience factory and the software process modeling approach from software engineering. CBR development experience is documented using software process models and stored in different levels of generality in a three-layered experience base. Up to now, experience from 9 industrial projects enacted by all INRECA II partners has been collected.

Automata-Theoretic vs. Property-Oriented Approaches for the Detection of Feature Interactions in IN
(1999)

The feature interaction problem in Intelligent Networks obstructs more and morethe rapid introduction of new features. Detecting such feature interactions turns out to be a big problem. The size of the systems and the sheer computational com-plexity prevents the system developer from checking manually any feature against any other feature. We give an overview on current (verification) approaches and categorize them into property-oriented and automata-theoretic approaches. A comparisonturns out that each approach complements the other in a certain sense. We proposeto apply both approaches together in order to solve the feature interaction problem.

This paper presents a completely systematic design procedure for asynchronous controllers.The initial step is the construction of a signal transition graph (STG, an interpreted Petri net) ofthe dialog between data path and controller: a formal representation without reference to timeor internal states. To implement concurrently operating control structures, and also to reducedesign effort and circuit cost, this STG can be decomposed into overlapping subnets. A univer-sal initial solution is then obtained by algorithmically constructing a primitive flow table fromeach component net. This step links the procedure to classical asynchronous design, in particu-lar to its proven optimization methods, without restricting the set of solutions. In contrast toother approaches, there is no need to extend the original STG intuitively.

An interrupter for use in a daisy-chained VME bus interrupt system has beendesigned and implemented as an asynchronous sequential circuit. The concur-rency of the processes posed a design problem that was solved by means of asystematic design procedure that uses Petri nets for specifying system and in-terrupter behaviour, and for deriving a primitive flow table. Classical designand additional measures to cope with non-fundamental mode operation yieldeda coded state-machine representation. This was implemented on a GAL 22V10,chosen for its hazard-preventing structure and for rapid prototyping in studentlaboratories.

Location problems with Q (in general conflicting) criteria are considered. After reviewing previous results of the authors dealing with lexicographic and Pareto location the main focus of the paper is on max-ordering locations. In these location problems the worst of the single objectives is minimized. After discussing some general results (including reductions to single criterion problems and the relation to lexicographic and Pareto locations) three solution techniques are introduced and exemplified using one location problem class, each: The direct approach, the decision space approach and the objective space approach. In the resulting solution algorithms emphasis is on the representation of the underlying geometric idea without fully exploring the computational complexity issue. A further specialization of max-ordering locations is obtained by introducing lexicographic max-ordering locations, which can be found efficiently. The paper is concluded by some ideas about future research topics related to max-ordering location problems.

In this paper we deal with locating a line in the plane. If d is a distance measure our objective is to find a straight line l which minimizes f(l) of g(l) (see the paper for the definition of these functions). We show that for all distance measures d derived from norms, one of the lines minimizing f(l) contains at least two of the existing facilities. For the center objective we always get an optimal line which is at maximum distance from at least three of the existing facilities. If all weights are equal, there is an optimal line which is parallel to one facet of the convex hull of the existing facilities.

In this paper relationships between Pareto points and saddle points in multiple objective programming are investigated. Convex and nonconvex problems are considered and the equivalence between Pareto points and saddle points is proved in both cases. The results are based on scalarizations of multiple objective programs and related linear and augmented Lagrangian functions. Partitions of the index sets of objectives and constranints are introduced to reduce the size of the problems. The relevance of the results in the context of decision making is also discussed.

Discrete Decision Problems, Multiple Criteria Optimization Classes and Lexicographic Max-Ordering
(1999)

The topic of this paper are discrete decision problems with multiple criteria. We first define discrete multiple criteria decision problems and introduce a classification scheme for multiple criteria optimization problems. To do so we use multiople criteria optimization classes. The main result is a characterization of the class of lexicographic max-ordering problems by two very useful properties, reduction and regularity. Subsequently we discuss the assumptions under which the application of this specific MCO class is justified. Finally we provide (simple) solution methods to find optimal decisions in the case of discrete multiple criteria optimization problems.

In line location problems the objective is to find a straight line which minimizes the sum of distances, or the maximum distance, respectively to a given set of existing facilities in the plane. These problems have well solved. In this paper we deal with restricted line location problems, i.e. we have given a set in the plane where the line is not allowed to pass through. With the help of a geometric duality we solve such problems for the vertical distance and then extend these results to block norms and some of them even to arbitrary norms. For all norms we give a finite candidate set for the optimal line.

In this survey we deal with the location of hyperplanes in n-dimensional normed spaces, i.e., we present all known results and a unifying approach to the so-called median hyperplane problem in Minkowski spaces. We describe how to find a hyperplane H minimizing the weighted sum f(H) of distances to a given, finite set of demand points. In robust statistics and operations research such an optimal hyperplane is called a median hyperplane.After summarizing the known results for the Euclidean and rectangular situation, we show that for all distance measures d derived from norms one of the hyperplanes minimizing f(H) is the affine hull of n of the demand points and, moreover, that each median hyperplane is a halving one (in a sense defined below) with respect to the geiven point set. Also an independence of norm result for finding optimal hyperplanes with fixed slope will be given. Furthermore we discuss how these geometric criteria can be used for algorithmical approaches to median hyperplanes, with an extra discussion for the case of polyhedral norms. And finally a characterizatio of all smooth norms by a sharpened incidence criterion for median hyperplanes is mentioned.

A large set of criteria to evaluate formal methods for reactive systems is presented. To make this set more comprehensible, it is structured according to a Concept-Model of formal methods. It is made clear that it is necessary to make the catalogue more specific before applying it. Some of the steps needed to do so are explained. As an example the catalogue is applied within the context of the application domain building automation systems to three different formal methods: SDL, statecharts, and a temporallogic.

Today, the worlds and terminologies of mechanical engineering and software engineering coexist, but they do not always work together seamlessly. Both worlds have developed their own separate formal vocabulary for expressing their concepts as well as for capturing and communicating their respective domain knowledge. But, these two vocabularies are not unified, interwoven, or at least interconnected in a reasonable manner. Thus, the subject of this paper is a comparison of the vocabularies of the two fields, namely feature technology from the area of mechanical engineering and software design patterns from the software engineering domain. Therefore, a certain amount of definitions, history, examples, etc. is presented for features as well as for design patterns. After this, an analysis is carried out to identify analogies and differences. The main intention of this paper is to inform both worlds - mechanical and software engineering - about the other side's terminology and to start a discussion about potential mutual benefits and possibilities to bridge the gap between these two worlds, e.g. to improve the manageability of CAx product development processes.

In this paper we prove a reduction result for the number of criteria in convex multiobjective optimization. This result states that to decide wheter a point x in the decision space is pareto optimal it suffices to consider at most n? criteria at a time, where n is the dimension of the decision space. The main theorem is based on a geometric characterization of pareto, strict pareto and weak pareto solutions

Ramsey Numbers of K_m versus (n,k)-graphs and the Local Density of Graphs not Containing a K_m
(1999)

In this paper generalized Ramsey numbers of complete graphs K_m versus the set langle ,n,k angle of (n,k)-graphs are investigated. The value of r(K_m,langle n,k angle) is given in general for (relative to n) values of k small compared to n using a correlation with Turan numbers. These generalized Ramsey numbers con be used to determine the local densities of graphs not containing a subgraph K_m.

The Weber problem for a given finite set of existing facilities {cal E}x = {Ex_1,Ex_2, ... ,Ex_M} subset R^2 with positive weights w_m (m = 1, ... ,M) is to find a new facility X* in R^2 such that sum_{m=1}^{M} w_{m}d(X,Ex_m) is minimized for some distance function d. In this paper we consider distances defined by polyhedral gauges. A variation of this problem is obtained if barriers are introduced which are convex polygonal subsets of the plane where neither location of new facilities nor traveling is allowed. Such barriers like lakes, military regions, national parks or mountains are frequently encountered in practice.From a mathematical point of view barrier problems are difficult, since the prensence of barriers destroys the convexity of the objective function. Nevertheless, this paper establishes a descretization result: One of the grid points in the grid defined by the existing facilities and the fuundamental directions of the gauge distances can be proved to be an optimal location. Thus the barrier problem can be solved with a polynomial algorithm.

Kernel smoothing in nonparametric autoregressive schemes offers a powerful tool in modelling time series. In this paper it is shown that the bootstrap can be used for estimating the distribution of kernel smoothers. This can be done by mimicking the stochastic nature of the whole process in the bootstrap resampling or by generating a simple regression model. Consistency of these bootstrap procedures will be shown.

In this paper we consider generalizations of multifacility location problems in which as an additional constraint the new facilities are not allowed to be located in a presprcified region. We propose several different solution schemes for this non-convex optimization problem. These include a linear programming type approach, penalty approaches and barrier approaches. Moreover, structural results as well as illustratrive examples showing the difficulties of this problem are presented

To present the decision maker's (DM) preferences in multicriteria decision problems as a partially ordered set is an effective method to catch the DM's purpose and avoid misleading results. Since our paper is focused on minimal path problems, we regard the ordered set of edges (E,=). Minimal paths are defined in repect to power-ordered sets which provides an essential tool to solve such problems. An algorithm to detect minimal paths on a multicriteria minimal path problem is presented

Let P be a probability measure of the real line R such that each of the product measures P^{otimes n} assigns the value 1/2 to every half space in R^{n} having the origin as a boundary point. Then P is symmetric.Example: A strictly stable law on R is symmetric iff it has median zero. The treated symmetry problem is related to the problem of characterizing the distribution of X_1 by the distribution of (X_2 + X_1, ... ,X_n + X_1), with X_1, ... ,X_n being independent and identically distributed random variables.

In continous location problems we are given a set of existing facilities and we are looking for the location of one or several new facilities. In the classical approaches weights are assigned to existing facilities expressing the importance of the new facilities for the existing ones. In this paper, we consider a pointwise defined objective function where the weights are assigned to the existing facilities depending on the location of the new facility. This approach is shown to be a generalization of the median, center and centdian objective functions. In addition, this approach allows to formulate completely new location models. Efficient algorithms as well as structure results for this algebraic approach for location problems are presented. Extensions to the multifacility and restricted case are also considered.

In this paper we consider the problem of optimizing a piecewise-linear objective function over a non-convex domain. In particular we do not allow the solution to lie in the interior of a prespecified region R. We discuss the geometrical properties of this problems and present algorithms based on combinatorial arguments. In addition we show how we can construct quite complicated shaped sets R while maintaining the combinatorial properties.

In this paper we deal with the determination of the whole set of Pareto-solutions of location problems with respect to Q general criteria. These criteria include as particular instances median, center or cent-dian objective functions. The paper characterizes the set of Pareto-solutions of all these multicriteria problems. An efficient algorithm for the planar case is developed and its complexity is established. the proposed approach is more general than the previously published approaches to multicriteria location problems and includes almost all of them as particular instances.

The computational complexity of combinatorial multiple objective programming problems is investigated. NP-completeness and #P-completeness results are presented. Using two definitions of approximability, general results are presented, which outline limits for approximation algorithms. The performance of the well known tree and Christofides' heuristics for the TSP is investigated in the multicriteria case with respect to the two definitions of approximability.

Convex Operators in Vector Optimization: Directional Derivatives and the Cone of Decrease Directions
(1999)

The paper is devoted to the investigation of directional derivatives and the cone of decrease directions for convex operators on Banach spaces. We prove a condition for the existence of directional derivatives which does not assume regularity of the ordering cone K. This result is then used to prove that for continuous convex operators the cone of decrease directions can be represented in terms of the directional derivatices . Decrease directions are those for which the directional derivative lies in the negative interior of the ordering cone K. Finally, we show that the continuity of the convex operator can be replaced by its K-boundedness.

The notion of the balance number introduced in [3,page 139] through a certain set contraction procedure for nonscalarized multiobjective global optimization is represented via a min-max operation on the data of the problem. This representation yields a different computational procedure for the calculation of the balance number and allows us to generalize the approach for problems with countably many performance criteria.

In this paper we consider the problem of locating one new facility in the plane with respect to a given set of existing facility where a set of polygonal barriers restricts traveling. This non-convex optimization problem can be reduced to a finite set of convex subproblems if the objective function is a convex function of the travel distances between the new and the existing facilities (like e.g. the Median and Center objective functions). An exact Algorithm and a heuristic solution procedure based on this reduction result are developed.

Let rC and rD be two convexdistance funtions in the plane with convex unit balls C and D. Given two points, p and q, we investigate the bisector, B(p,q), of p and q, where distance from p is measured by rC and distance from q by rD. We provide the following results. B(p,q) may consist of many connected components whose precise number can be derived from the intersection of the unit balls, C nd D. The bisector can contain bounded or unbounded 2-dimensional areas. Even more surprising, pieces of the bisector may appear inside the region of all points closer to p than to q. If C and D are convex polygons over m and m vertices, respectively, the bisector B(p,q) can consist of at most min(m,n) connected components which contain at most 2(m+n) vertices altogether. The former bound is tight, the latter is tight up to an additive constant. We also present an optimal O(m+n) time algorithm for computing the bisector.

In planar location problems with barriers one considers regions which are forbidden for the siting of new facilities as well as for trespassing. These problems areimportant since they reflect various real-world situations.The resulting mathematical models have a non-convex objectivefunction and are therefore difficult to tackle using standardmethods of location theory even in the case of simple barriershapes and distance funtions.For the case of center objectives with barrier distancesobtained from the rectilinear or Manhattan metric it is shown that the problem can be solved by identifying a finitedominating set (FDS) the cardinality of which is bounded bya polynomial in the size of the problem input. The resultinggenuinely polynomial algorithm can be combined with bound computations which are derived from solving closely connectedrestricted location and network location problems.It is shown that the results can be extended to barrier center problems with respect to arbitrary block norms having fourfundamental directions.

In this paper we deal with an NP-hard combinatorial optimization problem, the k-cardinality tree problem in node weighted graphs. This problem has several applications , which justify the need for efficient methods to obtain good solutions. We review existing literature on the problem. Then we prove that under the condition that the graph contains exactly one trough, the problem can be solved in ploynomial time. For the general NP-hard problem we implemented several local search methods to obtain heuristics solutions, which are qualitatively better than solutions found by constructive heuristics and which require significantly less time than needed to obtain optimal solutions. We used the well known concepts of genetic algorithms and tabu search with useful extensions. We show that all the methods find optimal solutions for the class of graphs containing exactly one trough. The general performance of our methods as compared to other heuristics is illustrated by numerical results.

Given a finite set of points in the plane and a forbidden region R, we want to find a point X not an element of int(R), such that the weighted sum to all given points is minimized. This location problem is a variant of the well-known Weber Problem, where we measure the distance by polyhedral gauges and allow each of the weights to be positive or negative. The unit ball of a polyhedral gauge may be any convex polyhedron containing the origin. This large class of distance functions allows very general (practical) settings - such as asymmetry - to be modeled. Each given point is allowed to have its own gauge and the forbidden region R enables us to include negative information in the model. Additionally the use of negative and positive weights allows to include the level of attraction or dislikeness of a new facility. Polynomial algorithms and structural properties for this global optimization problem (d.c. objective function and a non-convex feasible set) based on combinatorial and geometrical methods are presented.

In this paper a new trend is introduced into the field of multicriteria location problems. We combine the robustness approach using the minmax regret criterion together with Pareto-optimality. We consider the multicriteria Weber location problem which consists of simultaneously minimizing a number of weighted sum-distance functions and the set of Pareto-optimal locations as its solution concept. For this problem, we characterize the Pareto-optimal solutions within the set of robust locations for the original weighted sum-distance functions. These locations have both the properties of stability and non-domination which are required in robust and multicriteria programming.

We present a mathematical knowledge base containing the factual know-ledge of the first of three parts of a textbook on semi-groups and automata,namely "P. Deussen: Halbgruppen und Automaten". Like almost all math-ematical textbooks this textbook is not self-contained, but there are somealgebraic and set-theoretical concepts not being explained. These concepts areadded to the knowledge base. Furthermore there is knowledge about the nat-ural numbers, which is formalized following the first paragraph of "E. Landau:Grundlagen der Analysis".The data base is written in a sorted higher-order logic, a variant of POST ,the working language of the proof development environment OmegaGamma mkrp. We dis-tinguish three different types of knowledge: axioms, definitions, and theorems.Up to now, there are only 2 axioms (natural numbers and cardinality), 149definitions (like that for a semi-group), and 165 theorems. The consistency ofsuch knowledge bases cannot be proved in general, but inconsistencies may beimported only by the axioms. Definitions and theorems should not lead to anyinconsistency since definitions form conservative extensions and theorems areproved to be consequences.

The paper shows that characterizing the causal relationship between significant events is an important but non-trivial aspect for understanding the behavior of distributed programs. An introduction to the notion of causality and its relation to logical time is given; some fundamental results concerning the characterization of causality are pre- sented. Recent work on the detection of causal relationships in distributed computations is surveyed. The relative merits and limitations of the different approaches are discussed, and their general feasibility is analyzed.

The problem of finding an optimal location X* minimizing the maximum Euclidean distance to existing facilities is well solved by e.g. the Elzinga-Hearn algorithm. In practical situations X* will however often not be feasible. We therefore suggest in this note a polynomial algorithm which will find an optimal location X^F in a feasible subset F of the plane R^2

In the following, we discuss a procedure for interpolating a spatial-temporal stochastic process. We stick to a particular, moderately general model but the approach can be easily transered to other similar problems. The original data, which motivated this work, are measurements of gas concentrations (SO2, NO, O2) and several meteorological parameters (temperature, sun radiation, procipitation, wind speed etc.). These date have been and are still recorded twice every hour at several irregularly located places in the forests of the state Rheinland-Pfalz as part of a program monitoring the air pollution in the forests.

We consider a multiple objective linear program (MOLP) max{Cx|Ax = b,x in N_{0}^{n}} where C = (c_ij) is the p x n - matrix of p different objective functions z_i(x) = c_{i1}x_1 + ... + c_{in}x_n , i = 1,...,p and A is the m x n - matrix of a system of m linear equations a_{k1}x_1 + ... + a_{kn}x_n = b_k , k=1,...,m which form the set of constraints of the problem. All coefficients are assumed to be natural numbers or zero. The set M of admissable solutions {hat x} is an admissible solution such that there exists no other admissable solution x' with C{hat x} Cx'. The efficient solutions play the role of optimal solutions for the MOLP and it is our aim to determine the set of all efficient solutions

In this paper we give the definition of a solution concept in multicriteria combinatorial optimization. We show how Pareto, max-ordering and lexicographically optimal solutions can be incorporated in this framework. Furthermore we state some properties of lexicographic max-ordering solutions, which combine features of these three kinds of optimal solutions. Two of these properties, which are desirable from a decision maker" s point of view, are satisfied if and only of the solution concept is that of lexicographic max-ordering.

The Weber Problem for a given finite set of existing facilities {cal E}x = {Ex_1,Ex_2, ... ,Ex_M} subset R^2 with positive weights w_m (m = 1, ... ,M) is to find a new fcility X* such that sum_{m=1}^{M} w_{m}d(X,Ex_m) is minimized for some distance function d. A variation of this problem is obtained of the existing facilities are situated on two sides of a linear barrier. Such barriers like rivers, highways, borders or mountain ranges are frequently encountered in practice. Structural results as well as algorithms for this non-convex optimization problem depending on the distance function and on the number and location of passages through the barrier are presented. A reduction to convex optimization problems is used to derive efficient algorithms.

In this paper we introduce a new type of single facility location problems on networks which includes as special cases most of the classical criteria in the literature. Structural results as well as a finite dominationg set for the optimal locations are developed. Also the extension to the multi-facility case is discussed.

In this paper network location problems with several objectives are discussed, where every single objective is a classical median objective function. We will lock at the problem of finding Pareto optimal locations and lexicographically optimal locations. It is shown that for Pareto optimal locations in undirected networks no node dominance result can be shown. Structural results as well as efficient algorithms for these multi-criteria problems are developed. In the special case of a tree network a generalization of Goldman's dominance algorithm for finding Pareto locations is presented.

An approach to generating all efficient solutions of multiple objective programs with piecewise linear objective functions and linear constraints is presented. The approach is based on the decomposition of the feasible set into subsets, referred to as cells, so that the original problem reduces to a series of lenear multiple objective programs over the cells. The concepts of cell-efficiency and complex-efficiency are introduced and their relationship with efficiency is examined. A generic algorithm for finding efficent solutions is proposed. Applications in location theory as well as in worst case analysis are highlighted.

Facility location problems in the plane play an important role in mathematical programming. When looking for new locations in modeling real-word problems, we are often confronted with forbidden regions, that are not feasible for the placement of new locations. Furthermore these forbidden regions may habe complicated shapes. It may be more useful or even necessary to use approcimations of such forbidden regions when trying to solve location problems. In this paper we develop error bounds for the approximative solution of restricted planar location problems using the so called sandwich algorithm. The number of approximation steps required to achieve a specified error bound is analyzed. As examples of these approximation schemes, we discuss round norms and polyhedral norms. Also computational tests are included.

In this paper we deal with the location of hyperplanes in n-dimensional normed spaces. If d is a distance measure, our objective is to find a hyperplane H which minimizes f(H) = sum_{m=1}^{M} w_{m}d(x_m,H), where w_m ge 0 are non-negative weights, x_m in R^n, m=1, ... ,M demand points and d(x_m,H)=min_{z in H} d(x_m,z) is the distance from x_m to the hyperplane H. In robust statistics and operations research such an optimal hyperplane is called a median hyperplane. We show that for all distance measures d derived from norms, one of the hyperplanes minimizing f(H) is the affine hull of n of the demand points and, moreover, that each median hyperplane is (ina certain sense) a halving one with respect to the given point set.

There are several good reasons to introduce classification schemes for optimization problems including, for instance, the ability for concise problem statement opposed to verbal, often ambiguous, descriptions or simple data encoding and information retrieval in bibliographical information systems or software libraries. In some branches like scheduling and queuing theory classification is therefore a widely accepted and appreciated tool. The aim of this paper is to propose a 5-position classification which can be used to cover all location problems. We will provide a list of currentliy available symbols and indicate its usefulness in a - necessarily non-comprehensive - list of classical location problems. The classification scheme is in use since 1992 and has since proved to be useful in research, software development, classroom, and for overview articles.

We consider wavelet estimation of the time-dependent (evolutionary) power spectrum of a locally stationary time series. Allowing for departures from stationary proves useful for modelling, e.g., transient phenomena, quasi-oscillating behaviour or spectrum modulation. In our work wavelets are used to provide an adaptive local smoothing of a short-time periodogram in the time-freqeuncy plane. For this, in contrast to classical nonparametric (linear) approaches we use nonlinear thresholding of the empirical wavelet coefficients of the evolutionary spectrum. We show how these techniques allow for both adaptively reconstructing the local structure in the time-frequency plane and for denoising the resulting estimates. To this end a threshold choice is derived which is motivated by minimax properties w.r.t. the integrated mean squared error. Our approach is based on a 2-d orthogonal wavelet transform modified by using a cardinal Lagrange interpolation function on the finest scale. As an example, we apply our procedure to a time-varying spectrum motivated from mobile radio propagation.

In order to improve the quality of software systems and to set up a more effective process for their development, many attempts have been made in the field of software engineering. Reuse of existing knowledge is seen as a promising way to solve the outstanding problems in this field. In previous work we have integrated the design pattern concept with the formal design language SDL, resulting in a certain kind of pattern formalization. For the domain of communication systems we have also developed a pool of SDL patterns with an accompanying process model for pattern application. In this paper we present an extension that combines the SDL pattern approach with the experience base concept. This extension supports a systematic method for empirical evaluation and continuous improvement of the SDL pattern approach. Thereby the experience base serves as a repository necessary for effective reuse of the captured knowledge. A comprehensive usage scenario is described which shows the advantages of the combined approach. To demonstrate its feasibility, first results of a research case study are given.

The purpose of this expose is to explain the generic design of a customized communication subsystem. The expose addresses both functional and non-functional aspects. Starting point is a real-time requirement from the application area building automation. We show how this application requirement and some background information about the application area lead to a system architecture, a communication service, a protocol architecture and to the selection, adaptation, and composition of protocol functionalities. The reader will probably be surprised how much effort is necessary in order to implement the innocuous, innocent, inconspicuous looking application requirement. Formal description techniques (FDTs) will be used in all design phases.

Today's communication systems are typically structured into several layers, where each layer realizes a fixed set of protocol functionalities. These functionalities have been carefully chosen such that a wide range of applications can be supported and protocols work in a general environment of networks. However, due to evolving network technologies as well as increased and varying demands of modern applications general-purpose protocol stacks are not always adequate. To improve this situation new flexible communication architectures have been developed which enable the configuration of customized communication subsystems by composing a proper set of reusable building blocks. In particular, several approaches to automatic configuration of communication subsystems have been reported in the literature. This report gives an overview of theses approaches (F-CCS, Da CaPo, x-Kernel, and ADAPTIVE) and, in particular, defines a framework, which identifies common architectural issues and configuration tasks.

A new approach for modelling time that does not rely on the concept of a clock is proposed. In order to establish a notion of time, system behaviour is represented as a joint progression of multiple threads of control, which satisfies a certain set of axioms. We show that the clock-independent time model is related to the well-known concept of a global clock and argue that both approaches establish the same notion of time.

Due to the large variety of modern applications and evolving network technologies, a small number of general-purpose protocol stacks will no longer be sufficient. Rather, customization of communication protocols will play a major role. In this paper, we present an approach that has the potential to substantially reduce the effort for designing customized protocols. Our approach is based on the concept of design patterns, which is well-established in object oriented software development. We specialize this concept to communication protocols, and - in addition - use formal description techniques (FDTs) to specify protocol design patterns as well as rules for their instantiation and composition. The FDTs of our choice are SDL-92 and MSCs, which offer suitable language support. We propose an SDL pattern description template and relate pattern-based configuring of communication protocols to existing SDL methodologies. Particular SDL patterns and the configuring of a customized resource reservation protocol are presented in detail.

A non-trivial real-time requirement obeying a pattern that can be foundin various instantiations in the application domain building automation, and which is therefore called generic, is investigated in detail. Starting point is a description of a real-time problem in natural language augmented by a diagram, in a style often found in requirements documents. Step by step, this description is made more precise and finally transformed into a surprisingly concise formal specification, written in real-time temporal logic with customized operators. Wereason why this formal specification precisely captures the original description- as far as this is feasible due to the lack of precision of natural language.

A Tailored Real Time Temporal Logic for Specifying Requirements of Building Automation Systems
(1999)

A tailored real time temporal logic for specifying requirements of building automation systems is introduced and analyzed. The logic features several new real time operators, which are chosen with regard to the application area. The new operators improve the conciseness and readability of requirements as compared to a general-purpose real time temporal logic. In addition, some of the operators also enhance the expressiveness of the logic. A number of properties of the new operators are presented and proven.

A generic approach to the formal specification of system requirements is presented. It is based on a pool of requirement patterns, which are related to design patterns well-known in object-oriented software development. The application of such patterns enhances the reusability and genericity as well as the intelligibility of the formal requirement specification. The approach is instantiated by a tailored real-time temporal logic and by selecting building automation systems as application domain. With respect to this domain, the pattern discovery and reuse tasks are explained and illustrated, and a set of typical requirement patterns is presented. Finally, the results of a case study where the approach has been applied are summarized.

A straightforward formulation of a mathematical problem is mostly not ad-equate for resolution theorem proving. We present a method to optimize suchformulations by exploiting the variability of first-order logic. The optimizingtransformation is described as logic morphisms, whose operationalizations aretactics. The different behaviour of a resolution theorem prover for the sourceand target formulations is demonstrated by several examples. It is shown howtactical and resolution-style theorem proving can be combined.

We show how to buildup mathematical knowledge bases usingframes. We distinguish three differenttypes of knowledge: axioms, definitions(for introducing concepts like "set" or"group") and theorems (for relating theconcepts). The consistency of such know-ledge bases cannot be proved in gen-eral, but we can restrict the possibilit-ies where inconsistencies may be impor-ted to very few cases, namely to the oc-currence of axioms. Definitions and the-orems should not lead to any inconsisten-cies because definitions form conservativeextensions and theorems are proved to beconsequences.

In most cases higher-order logic is based on the (gamma)-calculus in order to avoid the infinite set of so-called comprehension axioms. However, there is a price to be paid, namelyan undecidable unification algorithm. If we do not use the(gamma) - calculus, but translate higher-order expressions intofirst-order expressions by standard translation techniques, we haveto translate the infinite set of comprehension axioms, too. Ofcourse, in general this is not practicable. Therefore such anapproach requires some restrictions such as the choice of thenecessary axioms by a human user or the restriction to certainproblem classes. This paper will show how the infinite class ofcomprehension axioms can be represented by a finite subclass,so that an automatic translation of finite higher-order prob-lems into finite first-order problems is possible. This trans-lation is sound and complete with respect to a Henkin-stylegeneral model semantics.

Extending existing calculi by sorts is astrong means for improving the deductive power offirst-order theorem provers. Since many mathemat-ical facts can be more easily expressed in higher-orderlogic - aside the greater power of higher-order logicin principle - , it is desirable to transfer the advant-ages of sorts in the first-order case to the higher-ordercase. One possible method for automating higher-order logic is the translation of problem formulationsinto first-order logic and the usage of first-order the-orem provers. For a certain class of problems thismethod can compete with proving theorems directlyin higher-order logic as for instance with the TPStheorem prover of Peter Andrews or with the Nuprlproof development environment of Robert Constable.There are translations from unsorted higher-order lo-gic based on Church's simple theory of types intomany-sorted first-order logic, which are sound andcomplete with respect to a Henkin-style general mod-els semantics. In this paper we extend correspond-ing translations to translations of order-sorted higher-order logic into order-sorted first-order logic, thus weare able to utilize corresponding first-order theoremprover for proving higher-order theorems. We do notuse any (lambda)-expressions, therefore we have to add so-called comprehension axioms, which a priori makethe procedure well-suited only for essentially first-order theorems. However, in practical applicationsof mathematics many theorems are essentially first-order and as it seems to be the case, the comprehen-sion axioms can be mastered too.

We tested the GYROSTAR ENV-05S. This device is a sensor for angular velocity. There- fore the orientation must be calculated by integration of the angular velocity over time. The devices output is a voltage proportional to the angular velocity and relative to a reference. The test where done to find out under which conditions it is possible to use this device for estimation of orientation.

A map for an autonomous mobile robot (AMR) in an indoor environment for the purpose ofcontinuous position and orientation estimation is discussed. Unlike many other approaches, this map is not based on geometrical primitives like lines and polygons. An algorithm is shown , where the sensordata of a laser range finder can be used to establish this map without a geometrical interpretation of the data. This is done by converting single laser radar scans to statistical representations of the environ-ment, so that a crosscorrelation of an actu al converted scan and this representative results into the actual position and orientation in a global coordinate system. The map itsel f is build of representative scansfor the positions where the AMR has been, so that it is able to find its position and orientation by c omparing the actual scan with a scan stored in the map.

One of the problems of autonomous mobile systems is the continuous tracking of position and orientation. In most cases, this problem is solved by dead reckoning, based on measurement of wheel rotations or step counts and step width. Unfortunately dead reckoning leads to accumulation of drift errors and is very sensitive against slippery. In this paper an algorithm for tracking position and orientation is presented being nearly independent from odometry and its problems with slippery. To achieve this results, a rotating range-finder is used, delivering scans of the environmental structure. The properties of this structure are used to match the scans from different locations in order to find their translational and rotational displacement. For this purpose derivatives of range-finder scans are calculated which can be used to find position and orientation by crosscorrelation.

In this paper we generalize the notion of method for proofplanning. While we adopt the general structure of methods introducedby Alan Bundy, we make an essential advancement in that we strictlyseparate the declarative knowledge from the procedural knowledge. Thischange of paradigm not only leads to representations easier to under-stand, it also enables modeling the important activity of formulatingmeta-methods, that is, operators that adapt the declarative part of exist-ing methods to suit novel situations. Thus this change of representationleads to a considerably strengthened planning mechanism.After presenting our declarative approach towards methods we describethe basic proof planning process with these. Then we define the notion ofmeta-method, provide an overview of practical examples and illustratehow meta-methods can be integrated into the planning process.

We argue in this paper that sophisticated mi-croplanning techniques are required even formathematical proofs, in contrast to the beliefthat mathematical texts are only schematicand mechanical. We demonstrate why para-phrasing and aggregation significantly en-hance the flexibility and the coherence ofthe text produced. To this end, we adoptedthe Text Structure of Meteer as our basicrepresentation. The type checking mecha-nism of Text Structure allows us to achieveparaphrasing by building comparable combi-nations of linguistic resources. Specified interms of concepts in an uniform ontologicalstructure called the Upper Model, our se-mantic aggregation rules are more compactthan similar rules reported in the literature.

Extending the planADbased paradigm for auto-mated theorem proving, we developed in previ-ous work a declarative approach towards rep-resenting methods in a proof planning frame-work to support their mechanical modification.This paper presents a detailed study of a classof particular methods, embodying variations ofa mathematical technique called diagonaliza-tion. The purpose of this paper is mainly two-fold. First we demonstrate that typical math-ematical methods can be represented in ourframework in a natural way. Second we illus-trate our philosophy of proof planning: besidesplanning with a fixed repertoire of methods,metaADmethods create new methods by modify-ing existing ones. With the help of three differ-ent diagonalization problems we present an ex-ample trace protocol of the evolution of meth-ods: an initial method is extracted from a par-ticular successful proof. This initial method isthen reformulated for the subsequent problems,and more general methods can be obtained byabstracting existing methods. Finally we comeup with a fairly abstract method capable ofdealing with all the three problems, since it cap-tures the very key idea of diagonalization.

Most automated theorem provers suffer from the problem thatthey can produce proofs only in formalisms difficult to understand even forexperienced mathematicians. Effort has been made to reconstruct naturaldeduction (ND) proofs from such machine generated proofs. Although thesingle steps in ND proofs are easy to understand, the entire proof is usuallyat a low level of abstraction, containing too many tedious steps. To obtainproofs similar to those found in mathematical textbooks, we propose a newformalism, called ND style proofs at the assertion level , where derivationsare mostly justified by the application of a definition or a theorem. Aftercharacterizing the structure of compound ND proof segments allowing asser-tion level justification, we show that the same derivations can be achieved bydomain-specific inference rules as well. Furthermore, these rules can be rep-resented compactly in a tree structure. Finally, we describe a system calledPROVERB , which substantially shortens ND proofs by abstracting them tothe assertion level and then transforms them into natural language.

Planning Argumentative Texts
(1999)

This paper presents PROVERB a text planner forargumentative texts. PROVERB's main feature isthat it combines global hierarchical planning and un-planned organization of text with respect to local de-rivation relations in a complementary way. The formersplits the task of presenting a particular proof intosubtasks of presenting subproofs. The latter simulateshow the next intermediate conclusion to be presentedis chosen under the guidance of the local focus.

This paper outlines an implemented system called PROVERB that explains machine -found natural deduction proofs in natural language. Different from earlier works, we pursue a reconstructive approach. Based on the observation that natural deduction proofs are at a too low level of abstraction compared with proofs found in mathematical textbooks, we define first the concept of so-called assertion level inference rules. Derivations justified by these rules can intuitively be understood as the application of a definition or a theorem. Then an algorithm is introduced that abstracts machine-found ND proofs using the assertion level inference rules. Abstracted proofs are then verbalized into natural language by a presentation module. The most significant feature of the presentation module is that it combines standard hierarchical text planning and techniques that locally organize argumentative texts based on the derivation relation under the guidance of a focus mechanism. The behavior of the system is demonstrated with the help of a concrete example throughout the paper.

We describe a technique to make application programs fault tolerant. This techADnique is based on the concept of checkpointing from an active program to one ormore passive backup copies which serve as an abstraction of stable memory. Ifthe primary copy fails, one of the backup copies takes over and resumes processADing service requests. After each failure a new backup copy is created in order torestore the replication degree of the service. All mechanisms necessary to achieveand maintain fault tolerance can be added automatically to the code of a nonADfaulttolerant server, thus making fault tolerance completely transparent for the applicaADtion programmer.

This paper deals with the reference choices involved in thegeneration of argumentative text. A piece of argument-ative text such as the proof of a mathematical theoremconveys a sequence of derivations. For each step of de-rivation, the premises (previously conveyed intermediateresults) and the inference method (such as the applica-tion of a particular theorem or definition) must be madeclear. The appropriateness of these references cruciallyaffects the quality of the text produced.Although not restricted to nominal phrases, our refer-ence decisions are similar to those concerning nominalsubsequent referring expressions: they depend on theavailability of the object referred to within a context andare sensitive to its attentional hierarchy . In this paper,we show how the current context can be appropriatelysegmented into an attentional hierarchy by viewing textgeneration as a combination of planned and unplannedbehavior, and how the discourse theory of Reichmann canbe adapted to handle our special reference problem.

Most automated theorem provers suffer from the problemthat the resulting proofs are difficult to understand even for experiencedmathematicians. An effective communication between the system andits users, however, is crucial for many applications, such as in a mathematical assistant system. Therefore, efforts have been made to transformmachine generated proofs (e.g. resolution proofs) into natural deduction(ND) proofs. The state-of-the-art procedure of proof transformation fol-lows basically its completeness proof: the premises and the conclusionare decomposed into unit literals, then the theorem is derived by mul-tiple levels of proofs by contradiction. Indeterminism is introduced byheuristics that aim at the production of more elegant results. This inde-terministic character entails not only a complex search, but also leads tounpredictable results.In this paper we first study resolution proofs in terms of meaningful op-erations employed by human mathematicians, and thereby establish acorrespondence between resolution proofs and ND proofs at a more ab-stract level. Concretely, we show that if its unit initial clauses are CNFsof literal premises of a problem, a unit resolution corresponds directly toa well-structured ND proof segment that mathematicians intuitively un-derstand as the application of a definition or a theorem. The consequenceis twofold: First it enhances our intuitive understanding of resolutionproofs in terms of the vocabulary with which mathematicians talk aboutproofs. Second, the transformation process is now largely deterministicand therefore efficient. This determinism also guarantees the quality ofresulting proofs.

Even though it is not very often admitted, partial functionsdo play a significant role in many practical applications of deduction sys-tems. Kleene has already given a semantic account of partial functionsusing a three-valued logic decades ago, but there has not been a satisfact-ory mechanization. Recent years have seen a thorough investigation ofthe framework of many-valued truth-functional logics. However, strongKleene logic, where quantification is restricted and therefore not truth-functional, does not fit the framework directly. We solve this problemby applying recent methods from sorted logics. This paper presents atableau calculus that combines the proper treatment of partial functionswith the efficiency of sorted calculi.

The semantics of everyday language and the semanticsof its naive translation into classical first-order language consider-ably differ. An important discrepancy that is addressed in this paperis about the implicit assumption what exists. For instance, in thecase of universal quantification natural language uses restrictions andpresupposes that these restrictions are non-empty, while in classi-cal logic it is only assumed that the whole universe is non-empty.On the other hand, all constants mentioned in classical logic arepresupposed to exist, while it makes no problems to speak about hy-pothetical objects in everyday language. These problems have beendiscussed in philosophical logic and some adequate many-valuedlogics were developed to model these phenomena much better thanclassical first-order logic can do. An adequate calculus, however, hasnot yet been given. Recent years have seen a thorough investigationof the framework of many-valued truth-functional logics. UnfortuADnately, restricted quantifications are not truth-functional, hence theydo not fit the framework directly. We solve this problem by applyingrecent methods from sorted logics.

Even though it is not very often admitted, partial functionsdo play a significant role in many practical applications of deduction sys-tems. Kleene has already given a semantic account of partial functionsusing a three-valued logic decades ago. This approach allows rejectingcertain unwanted formulae as faulty, which the simpler two-valued onesaccept. We have developed resolution and tableau calculi for automatedtheorem proving that take the restrictions of the three-valued logic intoaccount, which however have the severe drawback that existing theo-rem provers cannot directly be adapted to the technique. Even recentlyimplemented calculi for many-valued logics are not well-suited, since inthose the quantification does not exclude the undefined element. In thiswork we show, that it is possible to enhance a two-valued theorem proverby a simple strategy so that it can be used to generate proofs for the the-orems of the three-valued setting. By this we are able to use an existingtheorem prover for a large fragment of the language.

Typical instances, that is, instances that are representative for a particular situ-ation or concept, play an important role in human knowledge representationand reasoning, in particular in analogical reasoning. This wellADknown obser-vation has been a motivation for investigations in cognitive psychology whichprovide a basis for our characterization of typical instances within conceptstructures and for a new inference rule for justified analogical reasoning withtypical instances. In a nutshell this paper suggests to augment the proposi-tional knowledge representation system by a non-propositional part consistingof concept structures which may have directly represented instances as ele-ments. The traditional reasoning system is extended by a rule for justifiedanalogical inference with typical instances using information extracted fromboth knowledge representation subsystems.

This paper addresses two modi of analogical reasoning. Thefirst modus is based on the explicit representation of the justificationfor the analogical inference. The second modus is based on the repre-sentation of typical instances by concept structures. The two kinds ofanalogical inferences rely on different forms of relevance knowledge thatcause non-monotonicity. While the uncertainty and non-monotonicity ofanalogical inferences is not questioned, a semantic characterization ofanalogical reasoning has not been given yet. We introduce a minimalmodel semantics for analogical inference with typical instances.

Dynamic Lambda Calculus
(1999)

The goal of this paper is to lay a logical foundation for discourse theories by providing analgebraic foundation of compositional formalisms for discourse semantics as an analogon tothe simply typed (lambda)-calculus. Just as that can be specialized to type theory by simply providinga special type for truth values and postulating the quantifiers and connectives as constantswith fixed semantics, the proposed dynamic (lambda)-calculus DLC can be specialized to (lambda)-DRT byessentially the same measures, yielding a much more principled and modular treatment of(lambda)-DRT than before; DLC is also expected to eventually provide a conceptually simple basisfor studying higher-order unification for compositional discourse theories.Over the past few years, there have been a series of attempts [Zee89, GS90, EK95, Mus96,KKP96, Kus96] to combine the Montagovian type theoretic framework [Mon74] with dynamicapproaches, such as DRT [Kam81]. The motivation for these developments is to obtain a generallogical framework for discourse semantics that combines compositionality and dynamic binding.Let us look at an example of compositional semantics construction in (lambda)-DRT which is one ofthe above formalisms [KKP96, Kus96]. By the use of fi-reduction we arrive at a first-order DRTrepresentation of the sentence A i man sleeps. (i denoting an index for anaphoric binding.)

Higher-Order Tableaux
(1999)

Even though higher-order calculi for automated theorem prov-ing are rather old, tableau calculi have not been investigated yet. Thispaper presents two free variable tableau calculi for higher-order logicthat use higher-order unification as the key inference procedure. Thesecalculi differ in the treatment of the substitutional properties of equival-ences. The first calculus is equivalent in deductive power to the machine-oriented higher-order refutation calculi known from the literature, whereasthe second is complete with respect to Henkin's general models.

Many mathematical proofs are hard to generate forhumans and even harder for automated theoremprovers. Classical techniques of automated theoremproving involve the application of basic rules, of built-in special procedures, or of tactics. Melis (Melis 1993)introduced a new method for analogical reasoning inautomated theorem proving. In this paper we showhow the derivational analogy replay method is relatedand extended to encompass analogy-driven proof planconstruction. The method is evaluated by showing theproof plan generation of the Pumping Lemma for con-text free languages derived by analogy with the proofplan of the Pumping Lemma for regular languages.This is an impressive evaluation test for the analogicalreasoning method applied to automated theorem prov-ing, as the automated proof of this Pumping Lemmais beyond the capabilities of any of the current auto-mated theorem provers.

This paper addresses the decomposition of proofs as a means of constructingmethods in plan-based automated theorem proving. It shows also, howdecomposition can beneficially be applied in theorem proving by analogy.Decomposition is also useful for human-style proof presentation. We proposeseveral decomposition techniques that were found to be useful in automatedtheorem proving and give examples of their application.

This paper analyzes how mathematicians prove the-orems. The analysis is based upon several empiricalsources such as reports of mathematicians and math-ematical proofs by analogy. In order to combine thestrength of traditional automated theorem provers withhuman-like capabilities, the questions arise: Whichproblem solving strategies are appropriate? Which rep-resentations have to be employed? As a result of ouranalysis, the following reasoning strategies are recog-nized: proof planning with partially instantiated meth-ods, structuring of proofs, the transfer of subproofs andof reformulated subproofs. We discuss the represent-ation of a component of these reasoning strategies, aswell as its properties. We find some mechanisms neededfor theorem proving by analogy, that are not providedby previous approaches to analogy. This leads us to acomputational representation of new components andprocedures for automated theorem proving systems.

This paper shows how a new approach to theorem provingby analogy is applicable to real maths problems. This approach worksat the level of proof-plans and employs reformulation that goes beyondsymbol mapping. The Heine-Borel theorem is a widely known result inmathematics. It is usually stated in R 1 and similar versions are also truein R 2 , in topology, and metric spaces. Its analogical transfer was proposedas a challenge example and could not be solved by previous approachesto theorem proving by analogy. We use a proof-plan of the Heine-Boreltheorem in R 1 as a guide in automatically producing a proof-plan of theHeine-Borel theorem in R 2 by analogy-driven proof-plan construction.

This paper addresses a model of analogy-driven theorem proving that is more general and cognitively more adequate than previous approaches. The model works at the level ofproof-plans. More precisely, we consider analogy as a control strategy in proof planning that employs a source proof-plan to guide the construction of a proof-plan for the target problem. Our approach includes a reformulation of the source proof-plan. This is in accordance with the well known fact that constructing ananalogy in maths often amounts to first finding the appropriate representation which brings out the similarity of two problems, i.e., finding the right concepts and the right level of abstraction. Several well known theorems were processed by our analogy-driven proof-plan construction that could not be proven analogically by previous approaches.

This paper addresses analogy-driven auto-mated theorem proving that employs a sourceproof-plan to guide the search for a proof-planof the target problem. The approach presen-ted uses reformulations that go beyond symbolmappings and that incorporate frequently usedre-representations and abstractions. Severalrealistic math examples were successfully pro-cessed by our analogy-driven proof-plan con-struction. One challenge example, a Heine-Borel theorem, is discussed here. For this ex-ample the reformulaitons are shown step bystep and the modifying actions are demon-strated.

Analogy in CLAM
(1999)

CL A M is a proof planner, developed by the Dream group in Edinburgh,that mainly operates for inductive proofs. This paper addresses the questionhow an analogy model that I developed independently of CL A M can beapplied to CL A M and it presents analogy-driven proof plan construction as acontrol strategy of CL A M . This strategy is realized as a derivational analogythat includes the reformulation of proof plans. The analogical replay checkswhether the reformulated justifications of the source plan methods hold inthe target as a permission to transfer the method to the target plan. SinceCL A M has very efficient heuristic search strategies, the main purpose ofthe analogy is to suggest lemmas, to replay not commonly loaded methods,to suggest induction variables and induction terms, and to override controlrather than to construct a target proof plan that can be built by CL A Mitself more efficiently.