Multi-objective Pareto genetic algorithms using fast elite updating
Guanqi Guo, Zhumei Tan, Guanci Yang · 2009
This paper investigates the multi-objective optimization Pareto genetic algorithms (MOPGA) for searching alternative non-dominated Pareto-optimal solutions. A kind of niching approach using clustering crowding and fast elite updating is designed to maintain population diversity and uniform distribution of non-dominated solutions. The time complexity analysis shows clustering crowding and fast elite updating is a cost-efficient niching method. The simulation optimization on various multi-objective 0/1 knapsack problems shows MOPGA is capable of approximating to Pareto front evenly and cost efficiently, and the convergence rate and the distribution uniformity are consistently superior to that of the strength Pareto evolutionary approach (SPEA).