自适应Tent混沌搜索的蚁狮优化算法

针对蚁狮算法存在的早熟收敛和不易得到全局最优解等问题,借鉴混沌优化算法,提出了自适应Tent混沌搜索蚁狮算法.该算法首先使用Tent混沌映射初始化种群,然后自适应调整混沌搜索空间得到最优解,改善适应度较差个体,提高种群整体的适应度和寻优效率,同时使用锦标赛策略选择蚁狮个体.最后,利用混沌算子优化蚂蚁随机游走行为,与蚁狮觅食行为形成了全局、局部并行搜索模式.分别使用复杂高维基准函数和航迹规划问题测试算法性能.其中,6个复杂高维基准函数的寻优测试实验表明,对于30维基准函数,该算法经过约0.5秒收敛到最优值;对于50维基准函数,约2秒收敛到最优值.与标准蚁狮算法和其他优化算法相比,该算法具有较好的收敛速度和寻优精度,适合复杂高维函数寻优.航迹规划实验表明,对于包含7个威胁源的空域环境,当搜索维度为10维时,该算法经过0.939秒,迭代30次基本可以达到航迹代价的全局最优值.与标准蚁狮算法相比,能够更加快速准确地得到一条满足要求的航迹,具有实际应用价值.

Ant lion optimizatopm algorithm based on self-adaptive Tent chaos search

To overcome premature convergence and local minimum of Ant Lion Optimizer (ALO) algorithm, an innovative ALO algorithm based on self-adaptive chaos search which is combined with Tent chaos algorithm is proposed. Tent mapping is applied to initialize individuals in the search space. Self-adaptive adjustment of chaos search scopes is presented to get the better solution, perfect the fitness of the worse individuals and enhance the whole individuals'' fitness and efficiency. At the same time, Tournament selection strategy is used to screen out ant lion individuals. At the end, ants'' random walking behavior is optimized by chaos operator to form a parallel search mode with ant lion foraging behavior, which is beneficial to the overall optimization performance to obtain the optima. Algorithm''s performance is tested through complex benchmark functions with high-dimension and path planning problem. Experimental results of the former further confirm that, as for benchmark functions with 30 dimensions, it costs about 0.5 s to converge to the optimum, and for 50 dimensions, it is about 2 s. Compared with ALO and other optimization algorithms, the improved algorithm not only accelerates the convergence rate and improves the precision, but it is also fit for complex high-dimensional functions optimization. Path planning test manifests that for airspace with 7 menaces, while the search dimension is 10, after 0.939 s, and 30 iterations, the global optimum of the path cost function can be got. Compared with ALO, it has advantage in speed and accuracy to get the specific path, and it is of great value in actual problems.