Guest editorial: Special issue on optimisation in multi-agent systems

Abstract

Multi-agent systems applications in a number of areas such as e-commerce, disaster man-agement, and information acquisition through embedded devices (e.g. wireless sensornetworks) have generated a number of new challenges for algorithm designers. These chal-lengesmainlytaketheformofveryhardoptimisationproblemsthataresubstantiallydifferentfrom problems traditionally dealt with in other areas (e.g. industrial processes or schedulingapplications).Morespecifically,novelchallengescomefromthedistributednatureofmulti-agentsystemswheretheactorsresideondifferentcomputationalunitsandcancommunicateonlyalimitedamountofinformationwiththeirneighbours.Moreover,theagentsmaybeact-ing on behalf of different stakeholders each with its own aims and objectives, have differentcomputation/communication capabilities, and be tied to physical devices prone to failures.Moreover, given the dynamic nature of the application scenarios, effective algorithms haveto provide anytime solutions and approximate techniques are often required/desirable.This special issue collects a selection of contributions that focus on optimisation tech-niques for multi-agent systems. Preliminary version of these contributions were presentedat the OptMAS 2009 and 2008 workshops, which were held in conjunction with the Inter-national Conference on Autonomous Agents and Multi-agent Systems. The workshops pro-vided a multi-disciplinary forum for multi-agent algorithm designers, complementing thewidespectrumofknowledgerequiredtosuccessfullyaddressthechallengesoutlinedabove.The contributions published in this special issue focus on several key aspects of thesechallenges and provide a range of perspectives on the common problem of providing effec-tive and efficient techniques to solve hard optimisation problems in a multi-agent context.Specifically, this issue includes the following contributions:– Benchmarking Hybrid Algorithms for Distributed Constraint Optimisation Games byA. C. Chapman, A. Rogers, and Nicholas R. Jennings—this contribution relates solutionalgorithms for Distributed Constraint Optimisation Problems using game theory. Specif-ically, DCOP techniques are analysed using a potential game characterisation and novelhybrid algorithms are proposed and empirically analysed.