智能优化算法的模块化测试与仿真
Modular Test and Simulation of Intelligent Optimization Algorithm
王苏彧 1瞿圆媛 1李庆玲1
作者信息
- 1. 中国矿业大学(北京)机电与信息工程学院,北京 100083
- 折叠
摘要
生活及工程应用中许多问题可归结为目标之间相互矛盾的多目标优化问题,其应用需求强烈、理论创新性强、模型建立难、求解方法复杂.提出了智能优化算法的模块化测试与仿真方法,以多目标粒子群算法为例,通过对算法本身参数配置,以及多测试函数、多算法综合对比测试,可得出算法求解的帕累托前沿以及解集的收敛性与分布性指标,进而可评判该算法对于求解同类优化问题时的性能优劣.将上述方法应用到煤矿井下掘进截割路径规划中,得出了高效能、强安全的掘进截割最优路径.智能优化算法可用于"人工智能"相关课程课堂教学中,分析求解多目标优化问题,也可用于实验教学演示、创新实验项目,有助于提高学生的科研素养与实践能力.
Abstract
Many problems in life and engineering applications can be attributed to multi-objective optimization problems with contradictory objectives,which have strong application requirements and strong theoretical innovation,and then the model establishment is difficult and the solution methods are complex.In this paper,a modular test and simulation method for intelligent optimization algorithms is proposed.Taking multi-objective particle swarm optimiza-tion as an example,through the parameter configuration experiments of the algorithm itself,as well as the comprehen-sive comparison experiments of multi-test functions and multi-algorithms,the Pareto of Fronts of the algorithm and the convergence and distribution indicators can be obtained.Then,the performance of the algorithm for solving similar optimization problems can be judged.Finally,the multi-objective optimization algorithm is applied to the planning of the tunneling cutting path in the underground coal mine,and the optimal cutting path with high efficiency and strong safety is obtained.Intelligent optimization algorithms can be used to analyze and solve multi-objective optimization problems in the classroom teaching of"artificial intelligence"related courses,and can also be used for experimental teaching demonstrations and innovative experimental projects,which is helpful to improve students·scientific research literacy and practical ability.
关键词
多目标优化/智能优化算法/多目标粒子群/路径规划Key words
Multi-objective optimization/Intelligent optimization algorithm/Multi-objective particle swarm/Path planning引用本文复制引用
基金项目
中国矿业大学(北京)教学改革项目(J210414)
国家自然科学基金(62003350)
出版年
2024
NETL
NSTL
万方数据