Multi-objective Differential Evolution Algorithm Based on Linear Physical Programming
Shen Yong-ping · Systems Engineering · 2011
A multi-objective differential evolution algorithm based on linear physical programming is proposed to solve the constrained multi-objective optimization problem.One critical mechanism embedded in this approach is applying linear physical programming to handle the constraints and to design the fitness function during differential evolution.Then the multi-objective nonlinear constrained optimization problem can be transformed into a single-objective nonlinear optimization problem.Once the designer's preferences are articulated,the corresponding optimal design can be obtained immediately.The decision-making process becomes easier and suites the realities of engineering design.The experimental result shows that the proposed approach can solve the multi-objective constrained optimization effectively.