基于多元优化算法的路径规划

A Path Planner Based on Multivariant Optimization Algorithm

本文提出了一种基于多元优化算法和贝塞尔曲线的启发式智能路径规划方法.该方法通过用贝塞尔曲线描述路径的方法把路径规划问题转化成最优化问题.然后,使用多元优化算法来寻找最优的贝塞尔曲线控制点以获得最优路径.多元优化算法智能搜素个体协同合作交替的对解空间进行全局、局部迭代搜索以找到最优解.多元优化算法的搜索个体(元)按照分工不同可以分为全局元和局部元.在一次迭代中,全局元首先探索整个解空间以找出更优的潜在解区域.然后,局部元在各个潜在解区域进行局部开采以改善解质量.可见,搜索元具有分工不同的多元化特点,多元优化算法也就因此而得名.分工不同的搜索元之间高效的沟通和合作保证了多元优化算法的良好性能.为了评估多元优化算法的性能,我们基于标准测试地图比较了多元优化算法与其它三种经典启发式智能路径规划算法.结果表明,我们提出的方法在最优性,稳定性和有效性上方面优于其它方法.

A heuristic intelligent path planning method based on the multivariant optimization algorithm and the Bezier curve is presented. The path planning problem is transformed into an optimization problem through using the Bezier curve to represent a path in this method. Then,the multivariant optimization algorithm is applied to find the optimal control points of the best Bezier curve,aiming at finding the optimal path. The multivariant optimization algorithm searches the solution space through iterations of alternative global and local search. According to the different responsibilities,the search individuals（ atoms） could be divided into two types： the global atoms and the local atoms. In each iteration,global atoms explore the whole solution space to local potential areas,and then,local atoms exploit each potential area. Obviously,atoms are characterized by multivariant responsibilities,hence the name of the multivariant optimization algorithm. The good performance of the multivariant optimization algorithm is ensured by the efficient communication and cooperation of multivariant atoms. To evaluate the performance of the multivariant optimization algorithm,comparative experiments against the other three classical heuristic path planning algorithms are carried out based on a standard testing map. The results showthat our proposed method is superior to the other methods in optimality,stability and efficiency.