激励机制改进蚁群优化算法用于全局路径规划

Incentive Mechanism Based Improvement of Ant Colony Optimization for Global Route Planning

为提高优化算法搜索能力,分析了基本蚁群优化算法和心理学家斯金纳的强化激励方法的基本原理,将正、负激励原理应用于改进基本蚁群优化算法,提出了基于激励机制的改进蚁群算法,并给出了其数学描述。将改进的算法应用于求解旅行商问题和避碰约束下的最短路径规划问题,并与基本算法进行比较。仿真试验显示,改进的蚁群算法有效搜索到最短路径,实现全局路径优化。由于采用了激励机制,使得种群中所有个体都能够积极向最优解移动,从而更快地找到最优解,其较之基本蚁群算法具有较快的收敛速度,整体性能优越,能够应用于求解路径规划等问题。

In order to improve search ability of the optimization algorithm, the basic principles of primary ant colony optimization （ACO） algorithm and reinforcement theory was proposed by psychologist B F Skinner, introduced positive and negative incentive mechanisms to improve the basic ACO algorithm, and proposed an incentive mechanism improved ant colony optimization （IM-ACO） with giving its mathematical description.The IM-ACO was used to solve traveling salesman problem （TSP） and path planning restricted by both shortest length and collision avoidance with obstacles, whose performances were compared with that of primary ACO.Simulation showed that the IM-ACO successfully achieved the optimal path, fulfilling the global optimization goals.Due to the incentive mechanism, individuals in ant population were able to actively move towards better solutions, and eventually fix the optimal solution more quickly.It was concluded that the IM-ACO lead a faster convergence speed and superior overall performance than the original algorithm did, which were suitable for solving path planning problems.