基于自适应和变游走方向的改进狼群算法

针对狼群算法涉及参数较多、步长参数无法动态调整、游走方向固定等缺点，提出了一种基于自适应和变游走方向的改进狼群算法.该算法改进了游走行为、召唤行为、围攻行为3个主要步骤的移动步长，特别是当游走行为的试探方向改进后，每头狼都能根据头狼位置的变化自动调节移动步长、更换游走方向，从而简化参数设定，提高收敛速度和求解精度.仿真结果表明，改进算法在低维单峰函数求解精度上较原算法有明显改善，亦进一步提高了高维多峰函数的求解精度.

Improved wolf pack algorithm based on adaptive step length and adjustable scouting direction

Noticing the shortcomings of Wolf Pack Algorithm(WPA), such as too many parameters, predetermined step length and fixed scouting direction, a wolf pack algorithm based on adaptive step length and regulable scouting direction, named Modified Adaptive and Changed Scouting Direction Wolf Pack Algorithm(MACWPA) is proposed. It makes modifications on the step length of the three major processes, namely scouting behavior, summoning behavior and beleaguering behavior, and adopts tentative direction of scouting behavior, providing wolf pack more artificial intelligence. Each wolf is able to adjust its step length as well as scouting direction according to the leader wolf's position, which simplifies parameter set up, accelerates the convergence speed and improves the optimization precision. Simulation results show that the optimization precision for low dimensional unimodal function is greatly improved by MACWPA compared with WPA. It also improves the optimization precision of high dimensional multimodal function.