DEA駆動の多目的粒子群最適化

抄録

The PSO algorithm uses the positions of personal best and global best to achieve results in single-objective optimization. To solve multi-objective optimization problems, this study introduces DEA to estimate Pareto Optimality and superior sets for each individual. With the Lagrange multiplier, we can calculate the target positions that may be close to the Pareto frontier, and use these positions as the global best for each individual. Then the efficiency of the current iteration result can be obtained by calculating the current best position of each individual with the position of the corresponding target on the Pareto Optimality. In this study, we employ the CCR model, a classic DEA model, and the Ranking method to solve the problem.

To demonstrate the effectiveness of the proposed method, we use several benchmark functions and modify the DEA-PSO model to suit the test function.