EMO'07 Proceedings of the 4th international conference on Evolutionary multi-criterion optimization
Soft Computing - A Fusion of Foundations, Methodologies and Applications
| Hi-index | 0.00 |
Developing efficient algorithms for dynamic constrained multi-objective optimization problems (DCMOPs) is very challenging. This paper describes an attraction based particle swarm optimization (PSO) algorithm with sphere search for such problems. A dynamic constrained multi-objective optimization problem is transformed into a series of static constrained multi-objective optimization problems by dividing the time period into several equal intervals. To speed up optimization process and reuse the information of Pareto optimal solutions obtained from previous time, a new method based on sphere search is proposed to generate the initial swarm for the next time interval. To deal with the transformed problem effectively, a new particle comparison strategy is proposed for handling constraints in the problem. A local search operator based on the concept of attraction is introduced for finding good search directions of the particles. The results show that the proposed algorithm can effectively track the varying Pareto fronts with time.