Controller Agents for Constraints Solving: Implementation and Use of CACS Prototype Sami AL-MAQTARI and Habib ABDULRAB LITIS – INSA of Rouen - France Place Émile Blondel - BP 8 76131 Mont-Saint-Aignan Cedex {sami.almaqtari and abdulrab}@insa-rouen.fr Abstract The Distributed Constraint Satisfaction Problem (DCSP) which is an emerged field from the integration between two paradigms of different nature: Multi- Agent Systems (MAS) that is characterized by the autonomy and the distribution of its entities and the Constraint Satisfaction Problem paradigm (CSP) where all constraints are treated in central manner as a black-box. In this paper we aim to propose a model called Controller-Agents for Constraints Solving or CACS for short to be used for solving DCSP. We aim in this paper to present a model called CACS for (Controller-Agent for Constraints Solving). A controller role is to encapsulate and verify some constraints assigned to it. This model allows grouping constraints to form a subset that will be treated together as a local problem inside the controller. Using this model allows also handling non-binary constraints easily and directly so that no translating of constraints into binary ones is needed. Based on CACS, a prototype of DCSP solver is built. This paper presents the implementation outlines of that prototype and its usage methodology. The prototype is built in Java using general interfaces of both MAS and CSP platforms. These interfaces allow users to use the platforms of their choice providing that they implement these interfaces with the chosen platforms. Keywords: Distributed systems, Multi-Agent System, Constraints satisfaction. 1. INTRODUCTION Many real life problems can be modeled using Constraint Programming paradigm. By using CSP, we describe a problem as a set of variables that can take their values from some domains and a set of constraints over these variables. These constraints are limits over the combinations that the variables values can take in the same time. Another interesting paradigm for modeling real systems is the Multi-Agent System paradigm. An agent represents an autonomous entity. A Multi-Agent System is a set of distributed agents interacting together in order to achieve a global goal. In the integration between these two different paradigms emerged the Distributed Constraint Satisfaction Problem (DCSP). Such a system can be viewed as a group of agents; each one of them owns some variables. These agents are connected between them by constraints that restrain the values variables can take in the same time. An agent may choose values for its variables, but these values should respect the inter-agents constraints. Algorithms like Asynchronous BackTracking (ABT) [1-3] are proposed to solve DCSP. Many variants of this algorithm are also introduced proposing more efficient ways to find solution(s) for a given DCSP. In this paper, we introduce the implementation of a prototype based on our algorithm for solving DCSP. The model is based on the idea of using a special type of agent in order to encapsulate inter-agent constraints. In our model we call such an agent a Controller Agent. In a Controller-Agent for Constraint Solving system (CACS), Controller Agents are distinguished from other agents or Variables’ Agents which contain variables. Controller Agents are responsible for validating variables values sent by Variables’ Agents. Among the advantages of this model is the possibility of processing non-binary constraints easily and directly. In a classical ABT algorithm [2, 4], constraints are binary relations between agents’ variables. In order to treat non-binary constraints, constraint translating techniques [5, 6] like hidden variables encoding or dual encoding methods is to be used. Treating non-binary constraint directly is an 2008 20th IEEE International Conference on Tools with Artificial Intelligence 1082-3409/08 $25.00 © 2008 IEEE DOI 10.1109/ICTAI.2008.118 205 2008 20th IEEE International Conference on Tools with Artificial Intelligence 1082-3409/08 $25.00 © 2008 IEEE DOI 10.1109/ICTAI.2008.118 205 2008 20th IEEE International Conference on Tools with Artificial Intelligence 1082-3409/08 $25.00 © 2008 IEEE DOI 10.1109/ICTAI.2008.118 205