CUED Publications database

Agent Swarm Optimization: A platform to solve complex optimization problems

Montalvo, I and Izquierdo, J and Herrera, M and Pérez, R (2010) Agent Swarm Optimization: A platform to solve complex optimization problems. In: UNSPECIFIED.

Full text not available from this repository.


Agent Swarm Optimization (ASO) is a newly introduced abstract term to refer to a class of extensible algorithms devoted to solve complex optimization problems where classical techniques and most recent evolutionary algorithms get stranded. ASO stands for Agents Swarms Optimization. ASO is based on a multi-agent philosophy that crystallizes on a platform where different kinds of agents may interact and co-operate in the solution of a given optimization problem. In ASO such different evolutionary algorithms as PSO, ACO, GA, etc. may coexist contributing with their own agents, which are endowed with their specific characteristics, to the collective solution of the same problem. Taking advantage of parallel and distributed computation, allowing real-time enrolment of new agents, and raising the possibility for a human (the user) to become an active agent in the solution process make ASO a first-rate option to solve important decision-making problems in engineering, including the consideration of either one or multiple objectives. In this paper we present an ASO application developed for the optimal design of water distribution systems. In this specific problem, where the main objective is the sizing of the network elements, it can be seen that the introduction of a set of problem-specific rules regarding the agents' behaviours, considerably favours the search of solutions. The work includes solutions for two real-world water distribution networks where the role of ASO clearly improves the quality of the performed multiobjective optimization. The synergic co-operation between agents of different species and the use of general search algorithms based on parallel and distributed computation open up new possibilities to face complex engineering problems, as shown in this work. © 2010 Civil-Comp Press.

Item Type: Conference or Workshop Item (UNSPECIFIED)
Divisions: Div E > Manufacturing Systems
Depositing User: Cron Job
Date Deposited: 18 May 2020 20:02
Last Modified: 10 Dec 2020 08:07