In classical planning, the planner is given a concrete goal; it returns a plan for it or a failure message. In the latter case, the user can either quit or modify the goal. For many applications, it is more convenient to let the user provide a more elaborate specification consisting of constraints and preferences over possible goal states. Then, let the system discover a plan for the most desirable among the feasible goal states. To materialize such an approach we require a formalism for specifying preferences and constraints over goals and an algorithm for solving the resulting constrained optimization problem. In this work we motivate the need for planning with preferences and constraints, suggest a rich, yet intuitive formalism for representing goal preferences in the context of a deterministic action model, discuss some of its properties, propose an efficient algorithm for planning with preferences and constraints based on this formalism, and provide extensive experimental analysis in an interesting new domain of configuration planning.

We examine whether and how the Multiagent Plan Coordination Problem, the problem of resolving interactions between the plans of multiple agents, can be cast as a Distributed Constraint Optimization Problem (DCOP). We use ADOPT, a state-of-the-art DCOP solver that can solve DCOPs in an asynchronous, parallel manner using local communication between individual computational agents. We then demonstrate how we can take advantage of novel flaw-assignment strategies and plan coordination algorithms to significantly improve the performance of ADOPT on representative coordination problems. We close with a consideration of possible advances in framing our DCOP representation of the Multiagent

Constraint satisfaction techniques are commonly used for solving scheduling problems, still they are rare in AI planning. Although there are several attempts to apply constraint satisfaction for solving AI planning problems, these techniques never became predominant in planning; and they never reached the success of, for example, SATbased planners. In this paper we argue that existing constraint models for classical AI planning are not fully using the power of constraint satisfaction; thus we propose a reformulation, which significantly improves their efficiency.

Abstract. Action description languages, such as A and B [6], are expressive instruments introduced for formalizing planning domains and problems. The paper starts by proposing a methodology to encode an action language (with conditional effects and static causal laws), a slight variation of B, using Constraint Logic Programming over Finite Domains. The approach is then generalized to lift the use of constraints to the level of the action language itself. A prototype implementation has been developed, and the preliminary results are presented and discussed. 1

Tackling real-world problems often requires to take various types of constraints into account. Such constraint types range from simple numerical comparators to complex resources. This article describes how planning techniques can be integrated with general constraintsolving frameworks, like SAT, IP and CP. In many cases, the complete planning problem can be cast in these frameworks. 1

We deal with the search process of the Graph-Plan algorithm in this paper. We concentrate on the problem of finding supports for a sub-goal which arises during the search. We model the problem of finding supports as a constraint satisfaction problem in which arc-consistency is maintained. Contrary to other works on the similar topic, we do not model the whole planning problem as a CSP but only a small sub-problem within the standard solving process. Our model is based on dual views of the problem which are connected by channeling constraints. We performed experiments with several variants of propagation in the constraint model through channeling constraints. Experiments confirmed that the dual view of the problem enhanced with maintaining of arc-consistency is a good choice.

Abstract. We are dealing with solving planning problems by the GraphPlan algorithm. We concentrate on solving a problem of finding supporting actions for a goal. This problem arises as a sub-problem many times during search for a solution. We showed in the paper that the supports problem is NP-complete. In order to improve the solving process of the supports problems we proposed a new global consistency technique which we call projection consistency. We present a polynomial algorithm for enforcing projection consistency. The projection consistency was implemented within our experimental planning system which we used for empirical evaluation. The empirical tests showed improvements in order of magnitudes compared to the standard GraphPlan (both in time and number of constraint checks). A significant improvement was also reached compared to the recent similar technique based on maintaining of arc-consistency.

We consider qualitative simulation involving a finite set of qualitative relations in presence of complete knowledge about their interrelationship. We show how it can be naturally captured by means of constraints expressed in temporal logic and constraint satisfaction problems. The constraints relate at each stage the ‘past ’ of a simulation with its ‘future’. The benefit of this approach is that it readily leads to an implementation based on constraint technology that can be used to generate simulations and to answer queries about them. 1

Abstract. We deal with the search process of the GraphPlan algorithm in this paper. We concentrate on the problem of finding supports for a sub-goal which arises during the search. We model the problem of finding supports as a constraint satisfaction problem in which arcconsistency or singleton arc-consistency is maintained. Contrary to other works on the similar topic, we do not model the whole planning problem as a CSP but only a small sub-problem within the standard solving process. Our model is based on dual views of the problem which are connected by channeling constraints. We performed experiments with several variants of propagation in the constraint model through channeling constraints. Experiments confirmed that the dual view of the problem enhanced with maintaining of arc-consistency is a good choice. 1

Abstract. We are addressing the solving process over difficult AI problems such as planning and Boolean formula satisfaction in this paper. We propose a method for disentangling the structure of the problem hidden in its formulation. Namely we are viewing the problem as a graph with vertices end edges that we decompose into several complete sub graphs. The clique decomposition of the graph provides us with information about the structure of the problem which is then used for further reasoning. This reasoning is primarily targeted on early determining that a certain decision made during the search is unpromising. We implemented our new method and performed an experimental evaluation in planning with planning graphs and in Boolean formula satisfaction. The per-formed experiments showed order-of-magnitude improvements in comparison with other state-of-the-art approaches.