Modeling Multi-agent Logistic Process System using Hybrid Automata Ammar Mohammed and Ulrich Furbach Koblenz-Landau University, Computer Science Department, D-56070 Koblenz, Germany Abstract. Multi-agent systems are a widely accepted solution to handle complex problems. One application of multi-agent system is autonomous logistics. In au- tonomous logistic processes, potentially every element in a logistic supply chain is modeled as cooperating software agent. Thus, there exist modeling languages that are used to model such multi-agent systems. However, these modeling lan- guages do not allow verifying the properties of systems. Hybrid automata can be used to model hybrid systems by capturing both discrete and continuous changes of a system. Fortunately, hybrid automata are equipped with formal semantics that make formal methods possible to apply to them in order to prove certain properties of the specified systems. In this paper, we model multi-agent system behaviors in autonomous logistic processes using the concept of hybrid automata. With the help of model checking techniques, we can prove some properties of a modeled system before involving in the implementation of a system. 1 Introduction Software agents and multi-agent systems (MAS) are an approach to implementing au- tonomous and interacting software systems. An agent is an autonomous decision maker on behalf of some real world entity. Generally, agents are able to perceive from their environment with sensors, and to act with actuators. The agents choose their action as because of a reasoning process. In particular, agents in multi-agent system are able to communicate and coordinate with each other to fulfill tasks in cooperation or com- petition depending on their respective goals and abilities. This enables them to solve complex problems and tasks in a distributed way [9] . There are several approaches for modeling multi-agent system like [11], [12]. Among of them, Agent UML [1] is one of the widely accepted languages for modeling multi- agent system . This language is chosen by FIPA 1 association as an acceptable language to model interactions among agents. Unfortunately, although most multi-agent system modeling languages are clear to understand, they are not able to verify some properties of the modeled system, because there is no formal semantics of agent decision making. One important aspect of multi-agent systems is that the agents interact with a physi- cal environment. Such interactions typically consist of continuous changes of behaviors of an agent (e.g. a movement of an agent in logistic transportation, or an agent waiting for some events), as well as discrete changes of the behaviors. The previous scenario 1 www.fipa.org Mohammed A. and Furbach U. (2008). Modeling Multi-agent Logistic Process System using Hybrid Automata. In Proceedings of the 6th International Workshop on Modelling, Simulation, Verification and Validation of Enterprise Information Systems, pages 141-149 DOI: 10.5220/0001729501410149 Copyright c  SciTePress