In this work a new approach to parent selection based on Grid-value in multiobjective genetic algorithm is proposed. Here grid is used as a frame to determine the location of individuals in the objective space. Every solution inside the grid maintains an objective-rank vector and summation value. Summation value is the scalar fitness and used to discriminate individuals instead of Pareto-dominance relation. Since multiple solutions occupy same grid have same Summation-value, an adaptive selection mechanism is used in order to avoid duplicate selection and thereby enhancing spread of solution on the Pare-to front. The multi-objective genetic algorithm based on the proposed selection scheme is tested on problems of CEC09 competition. The algorithm has shown either comparable or good performance on few unconstrained test problems.
展开▼
