分布式约束优化方法研究进展

多agent系统作为分布式人工智能研究领域的重要分支,已被广泛应用于多个领域中复杂系统的建模.而分布式约束优化作为一种多agent系统求解的关键技术,已成为约束推理研究的热点.首先对其适用性进行分析,并基于对已有算法的研究,总结出采用该方法解决问题的基本流程,在此基础上,从解的质量保证、求解策略等角度对算法进行了完整的分类;其次,根据算法分类结果以及执行机制,对大量经典以及近年来的分布式约束优化算法进行了深入分析,并从通信、求解质量、求解效率等方面对典型算法进行了实验对比;最后,结合分布式约束优化技术的求解优势给出了分布式约束优化问题的实际应用特征,总结了目前存在的一些问题,并对下一步工作进行了展望.

Research Progress in Distributed Constraint Optimization Method

Multi agent system, one of important branches of distributed artificial intelligence, has been widely applied to modeling a serious of complex systems in diverse research fields. Significant research effort has sought to solve constraint programming with distributed constraint optimization which is a popular framework for multi agent system. The contributions of this research proceed from previous work in the following ways. First, based on the existing research, the applicability of distributed constraint optimization is analyzed, and general process of distributed constraint optimization algorithms is extracted. Second, a relatively complete classification of algorithms is provided from the perspective of quality assurance and solving strategies. Next, considering execution mechanism, a thorough analysis of a large number of classic algorithms proposed in recent years is carried out. Moreover, the experimental analysis of some typical algorithms with the metrics of communication, solution quality and efficiency is provided. Finally, combining the advantage of distributed constraint optimization technology, the application characteristics of distributed constraint optimization problem are proposed, and future work is discussed.