Abstract: The problem solving model is based on the formal description of the home service robot task planning. An improved cultural algorithm is proposed. The problem is solved by the mutually reinforcing between the belief space and population space of the algorithms. A unique encoding method is applied in this algorithm. The population uses genetic algorithms as a means of evolution. The belief space of the algorithms is constructed by a unique extract method to promote its evolution. It is proved that the algorithm applied in the simulation platform for home service robot contest is more effective.

Key words: home service robot simulation, mission planning, cultural algorithm

摘要： 通过对家庭服务机器人任务规划问题进行形式化描述，给出了问题的求解模型，提出了一种改进的文化算法，通过算法中信念空间和种群空间的相互联系和相互促进实现求解。算法采用独特的编码方式，其种群空间采用遗传算法作为进化手段，采用较为独特的信念提取方式构造算法的信念空间并促使其进化。将该算法用于家庭服务机器人大赛的仿真平台上，证明其有效性。

关键词: 家庭服务机器人仿真, 任务规划, 文化算法