Multi-objective Optimization Algorithm Inspired by Ecological Prey-predator Model

A novel evolutionary multi-objective optimization algorithm based on the prey-predator model is proposed. The algorithm adjusts the objectives' weights to ecosystem populations using periodically dynamic attributes of the model without manual interference so as to reduce the dependences on problem acknowledgements and find a pareto archive fast. Parameters selection Criterion of weights' adjustment strategy is given and experiments on several difficult multi-objective optimization problems are done. The experimental results proved that the algorithm is effective.