传统的集中式进化计算方法在处理大规模复杂优化问题时存在计算效率方面的瓶颈问题。本项目将分布式计算的强大并行计算能力与Memetic算法的全局和局部优化能力相结合，提出一种面向大规模优化问题的新型分布式Memetic算法。首先，基于对已有进化算法分布式模型的总结分析，并结合Memetic算法的搜索特性，提出Memetic算法分布式并行实现的新模型。其次，在Memetic算法的分布式模型下构建多个并行子种群，通过多种群并行执行多种局部搜索策略，同时在进化过程中交换优化信息并相应调整局部搜索策略，实现多种群的协同进化；以及通过对算法优化状态和执行环境的动态感知，实现多种群的自适应协作通讯机制，进一步提高分布式Memetic算法的并行优化效率。最终，基于Memetic算法的分布式模型和协作通讯策略的研究，实现面向大规模优化的高效分布式Memetic算法，并应用于大规模无线传感器网络部署问题。

Traditional centralized evolutionary algorithms (EAs) face the bottleneck problem of computing efficiency when dealing with large-scale and complex optimization problems. By combining the parallel computing power of distributed computing devices with the global and local search capability of memetic algorithms, this project proposes a novel distributed memetic algorithm for solving large-scale optimization problems. By making analysis and summary of the existing distributed model for EAs and the search characteristic of memetic algorithm, this project first proposes a new distributed parallel model for memetic algorithm. Based on the distributed model, multiple sub-populations are constructed and different sub-populations perform different local search strategies. These sub-populations realize cooperative coevolution by exchanging optimization information and accordingly adjust their local search strategies during the evolutionary process. Moreover, through the dynamic perception of the optimization states and environment, the cooperative communication strategy between different sub-populations is adjusted adaptively. As a result, the optimization efficiency of the distributed memetic algorithm can be further enhanced. This project eventually develops an efficient distributed memetic algorithm for large-scale optimization, and the proposed algorithm will be applied to solving large-scale wireless sensor network deployment problems.