V. Marik, V. Vyatkin, A.W. Colombo (Eds.): HoloMAS 2007, LNAI 4659, pp. 185–194, 2007.
© Springer-Verlag Berlin Heidelberg 2007
Metaheuristic Agent Teams for Job Shop Scheduling
Problems
Mehmet E. Aydin
University of Bedfordshire, Dept. of Computing and Information Systems,
Luton, UK
mehmet.aydin@beds.ac.uk
Abstract. This paper addresses and introduces an overview on various multi-
agent architectures applied to teams of metaheuristic agents for job shop
scheduling applications, whose developed and examined on distributed problem
solving environments. We reported a couple of topologies; ATEAM is a
centrally coordinating method, which provides very good results when well-
studied, on the other hand, architectures based on peer-to-peer technology
provide wider flexibility in implementing various fashions. The experiment-
tation for each targeted topology has revealed more details and attracts more
attentions.
1 Introduction
Metaheuristic agents are multi agent systems identified to describe teams of search
agents used to operate for optimisation. This type of multi agent systems is specific to
implementations of metaheuristics to solve large optimisation problems [16]. One of
the major problems with metaheuristics is that there is no guarantee provided to hit
optimum solutions within a reasonable time, as they usually provide with local
optimum, which may not be very desirable most of the time. One way to overcome
this problem is to diversify the search conducted with the heuristics.
Distributed problem solving is mainly expected to bring simplicity and reduction in
computational time. In addition, there is a hope to provide with more diversity so as to
obtain a reasonably useful solution. That is the case if a well studied multi agent
system can tackle multiple regions of the search space simultaneously. Multiple
independent runs of the algorithms, which offer distributing the systems over the
particular metaheuristic agents, have a potential of carrying out concurrent search
within search spaces. We have not come across any study focusing on the topologies
of multi agent systems implemented for job-shop-scheduling problems with
metaheuristics in an overall view point. That will close the gap that appears in head of
new researchers with bringing forward multiple approaches for the issue. Throughout
this paper, a couple of multi agent topologies applied to metaheuristic agent teams for
solving job shop scheduling are introduced. Each topology provides with different
benefits in tackling search and problem solving. The applications are developed to
solve the classical job shop scheduling problems, which are known as very hard