Abstract:
This paper proposes a Multi-Objective Artificial Physics Optimization(MOAPO) algorithm to solve multi-objective optimization problem. By adopting the method of setting random power for each object to get the global object, both the global best and worst objects fitness in multi-objective optimization problem are found, so that searching for Pareto optimal set of multi-objective optimization problem is implemented by using APO algorithm. Three benchmark functions are tested to compare the performance of MOAPO with MOPSO. The results show that MOAPO is effective for solving multi-objective problems with an excellent diversity.
Key words:
Artificial Physics Optimization(APO),
multi-objective optimization,
aggregating functions,
dynamic changing,
diversity
摘要: 提出一种解决多目标优化问题的多目标拟态物理学优化(MOAPO)算法。该算法利用为每个目标赋予随机权重的方法求得全局总目标,由此选取全局最好及最差适应值,并利用拟态物理学优化算法实现对Pareto最优解集的搜索。通过3个典型多目标优化测试函数对MOAPO和MOPSO进行比较,结果表明了MOAPO算法的有效性,特别是在保持解集分布性方面具有较好的性能。
关键词:
拟态物理学优化,
多目标优化,
聚集函数法,
动态变化,
分布性
CLC Number:
WANG Yan, CENG Jian-Chao. Artificial Physics Optimization Algorithm Solving Multi-objective Optimization Problem[J]. Computer Engineering, 2010, 36(20): 188-190.
王艳, 曾建潮. 解决多目标优化问题的拟态物理学优化算法[J]. 计算机工程, 2010, 36(20): 188-190.