In the current computing environment, the interactive behaviour of software is a significant feature requested by many users. The minimal perturbation problem is often highly demanded and, in a sense, it seems to be very close to the interactivity. In this paper we want to compare these two features and also summarize their main demands and impacts on the constraint satisfaction problem. We also suggest several improvements toward these features.

Many real-life problems are dynamic, with changes in the problem definition occurring after a solution to the initial formulation has been reached. The minimal perturbation problem incorporates these changes, along with the initial solution, as a new problem whose solution must be as close as possible to the solution of an initial problem.

Abstract. The paper presents an iterative forward search framework for solving constraint satisfaction and optimization problems. This framework combines ideas of local search, namely improving a solution by local steps, with principles of depth-first search, in particular extending a partial feasible assignment towards a solution. Within this framework, we also propose and study a conflict-based statistics and explanationbased arc consistency maintenance. To show the versatility of the proposed framework, the dynamic backtracking algorithm with maintaining arc consistency is presented as a special instance of the iterative forward search framework. The presented techniques are compared on random constraint satisfaction problems and a real-life lecture timetabling problem. 1

This paper 2 details the stages of building a substantial, carefully specified, fully tested and fully operational university and school timetabling system. This is reported as a case study in applying Constraint Satisfaction techniques. The emphasis is on the software engineering aspects of the problem. That is, Constraint Satisfaction problems are expressed in a language more familiar to the formal software engineering community. Moreover, this language is used to formulate domain constraints and heuristic information. In addition to that, the user's needs are looked at more closely. For instance, the system supplies indications useful for relaxing or reformulating the constraints of the problem when a solution satisfying these constraints is impossible to produce. This has a value in bringing Constraint Satisfaction one-step closer to formal specification, program verification and transformation.

Abstract. Formulation of many real-life problems evolves as the problem is being solved. These changes are typically initiated by a user intervention or by changes in the environment. In this paper, we propose a formal description of so called minimal perturbation problem that allows an “automated” modification of the (partial) solution when the problem formulation changes. Our model is defined for constraint satisfaction problems with emphasis put on finding a solution anytime even for over-constrained problems.

Timetabling problems are often solved based upon what is assumed to be complete information about the problem. However, users may later desire modifications to the generated timetable to adapt to changing requirements or increase their personal satisfaction with the solution [1,5,3]. Methods that allow corrections to an existing timetable