A parallel evolutionary approach to multi-objective optimization

Evolutionary algorithms have been used since the mid-eighties to solve complex single and multi-objective optimization problems. More recently the swarm intelligent approaches such as particle swarm optimization and ant colony optimization have been successfully used for multi-objective optimization. This paper proposes a new approach based on the generic generalized particle model (GF-GPM) for computing in parallel approximate efficient solutions for the distribution problem with multiple objectives. Unlike the swarm optimization approaches, GE-GPM is inspired by physical models of particle dynamics. We use mathematical formulations to describe or predict the properties and evolution of different states of the particles. In particular, according to "differential equation theory", we develop efficient optimization techniques for multi-objective problems. We also adopt methods of classical mechanics to tackle the problem of modeling the interaction among the particles. We show that GE-GPM, being inspired by classical mechanics, enables feasible multi-objective optimization in very large scales. The GE-GPM approach has a low computational complexity, which is crucial for the functioning of large-scale distribution problems.