Many AI synthesis problems such as planning, scheduling or design may be encoded in a constraint satisfaction problem (CSP). A CSP is typically defined as the problem of finding any consistent labeling for a fixed set of variables satisfying all given constraints between these variables. However, for many real tasks, the set of constraints to consider may evolve because of the environment or because of user interactions. The problem we consider here is the solution maintenance problem in such a dynamic CSP (DCSP). We propose a new class of constraint recording algorithms called Nogood Recording that may be used for solving both static and dynamic CSPs. It offers an interesting compromise, polynomially bounded in space, between an ATMS-like approach and the usual static constraint satisfaction algorithms. 1 Introduction The constraint satisfaction problem (CSP) model is widely used to represent and solve various AI related problems and provides fundamental tools in areas such as truth...

Many AI problems can be modeled as constraint satisfaction problems (CSP), but many of them are actually dynamic: the set of constraints to consider evolves because of the environment, the user or other agents in the framework of a distributed system. In this context, computing a new solution from scratch after each problem change is possible, but has two important drawbacks: inefficiency and instability of the successive solutions. In this paper, we propose a method for reusing any previous solution and producing a new one by local changes on the previous one. First we give the key idea and the corresponding algorithm. Then we establish its properties: termination, correctness and completeness. We show how it can be used to produce a solution, either from an empty assignment, or from any previous assignment and how it can be improved using filtering or learning methods, such as forward-checking or nogoodrecording. Experimental results related to efficiency and stability are given, wit...

Many AI problems can be modeled as constraint satisfaction problems (CSP), but many of them are actually dynamic: the set of constraints to consider evolves because of the environment, the user or the other agents in the framework of a distributed system. The notion of dynamic CSP (DCSP) has been introduced to represent them. In spite of its name, the recent Dynamic Backtracking algorithm proposed in (Ginsberg 1993) does not solve dynamic CSPs, but static ones. But its nogood recording and backtracking mechanisms are very interesting in the DCSP framework. In this paper, we propose an extension of this algorithm which provides the user with explanations in case of inconsistency and allows dynamic CSPs to be dealt with very efficiently. After presenting the Dynamic Backtracking algorithm, its extension and how to use it in the DCSP framework, we present and discuss some experimental results. Dynamic backtracking In (Ginsberg 1993), Matthew L. Ginsberg proposed a new algorithm for solv...

Introduction A multimedia document is defined as a set of objects from different media (text, image, video, audio) that are spatially and temporally organized and on which a navigational structure can be set. Such an entity can be rendered thanks to a presentation engine by means of the output channels of the computer (screen and speakers). Numerous works [5], [19] and even standards [26], [16] have been done for the definition of languages and formats of multimedia documents, largely focusing on the temporal dimension of documents. They allow the specification of the temporal composition of media objects either by absolute placements [14], by event-based approaches [16], by the use of a hierarchy of temporal operators [19], by constraint based definitions [5], [11], [12], [2] or a combination of some of these methods [26]. In constraint based environments, the author can describe the spatial and temporal organization of a document by setting constraints between basic or com

This paper introduces a model for Distributed Employee Timetabling Problems (DisETPs) and proposes a general architecture for solving DisETPs by using a Multi Agent System (MAS) paradigm. The architecture is composed of a set of autonomous software Scheduling Agents (SAs) that solve the Employee Timetabling Problems (ETP) for each department. Each agent has its own local ETP problem and its own goals. The Scheduling Agents must coordinate these goals with the other agents in order to achieve a solution for the whole organization that yields a better result with respect to the global targets. To achieve a coherent and consistent global solution, the SAs make use of a sophisticated negotiation protocol among scheduling agents that always ends in an agreement (not ensured to be optimal). The main functionalities of this protocol are agent to agent relation definition, a mechanism to approve a chain of Request for Changes and an electronic marketplace for bidding on preferred common time slots. Experimental analysis of the implemented Multi Agent System for the Soroka medical center is presented. The results of our study indicate that the proposed framework has the potential to reduce the cost of transportation for the nurses scheduled by the hospital. 2 1

Today, many universities are opting for modular degree programmes. Such modular courses provide greater exibility for students. However, such a system is naturally complex; modules may feature pre and co-requisites and may run over dierent periods of times, and have dierent credit values. General university requirements will need to be met by students to continue their studies. Add to this timetable constraints and the module selection process can be a daunting task. Students may further complicate the process by explicitly wanting to take or avoid modules. They may require a general overview to see what options are available to them, such as the dierent routes to a particular degree. The University of Glasgow currently has no automated process to help with this. This paper describes our eorts in applying constraint programming to this con guration problem. We show how we went about tackling the problem using the constraint programming language Choco. We present a small example problem, a constraint programming model of this problem, and describe how we deliver explanations. We then present an extension of this model to deal with dynamic problems, where variables and constraints can be activated as a result of decisions made by the user or search process. Throughout this study, our goal has been to keep it simple, attempting to show that an o-the-shelf constraint programming toolkit is up to the task.

The Dynamic Constraint Satisfaction Problem (DCSP) formalism has been gaining attention as a valuable and often necessary extension of the static CSP framework. Dynamic Constraint Satisfaction enables CSP techniques to be applied more extensively, since it can be applied in domains where the set of constraints and variables involved in the problem evolves with time. At the same time, the Case-Based Reasoning (CBR) community has been working on techniques by which to reuse existing solutions when solving new problems. We have observed that dynamic constraint satisfaction matches very closely the case-based reasoning process of case adaptation. These observations emerged from our previous work on combining CBR and CSP to achieve a constraintbased adaptation. This paper summarizes our previous results, describes the similarity of the challenges facing both DCSP and case adaptation, and shows how CSP and CBR can together begin to address these challenges.

Introduction Le cadre des problemes de satisfaction de contraintes ou CSP (pour Constraint Satisfaction Problems) a pour finalite l'expression et la resolution de problemes faisant intervenir des contraintes. De facon tres generale, une contrainte correspond a l'enonce d'une proprie te relative a certaines caracteristiques de differents objets : propriete physique (spatiotemporelle. ..) necessairement satisfaite par les objets consideres ou propriete desiree par l'utilisateur par exemple. Du fait de sa generalite, il est assez naturel que la notion de contrainte soit frequemment utilisee et ait recu beaucoup d'attention de la part des communautes de l'intelligence artificielle ou de la recherche operationnelle, toutes deux preoccupees par la representation et la resolution de problemes. Face aux illustres predecesseurs que sont la programmation lineaire (en nombres entiers) ou la logique propositionnelle, le cad

Many real world Timetabling Problems (TTPs) are composed of organizational parts that need to timetable their sta# in an independent way, while adhering to some global constraints. Later the departmental timetables are combined to yield a coherent, consistent solution. This last phase involves negotiations with the various agents and requests for changes in their own solutions.

Our case-based reasoning (CBR) integration with the constraint satisfaction problem (CSP) formalism has undergone several transformations on its journey from initial research idea to product-intent design. Both unexpected research results as well as interesting insights into the real-world applicability of the integrated methodology emerged as the integration was explored from alternative viewpoints. In this paper, the alternative viewpoints and the results that were enabled by these viewpoints are described.