摘要
基于计算智能“隐并行性”实现多任务优化(multi-task optimization,MTO),是当前研究的热点和前沿技术.与传统单任务优化算法相比,通过挖掘群体智能内在并行和内涵并行同时优化多个任务,可显著提高问题求解质量以及缩短任务求解时间.首先,对MTO相关英文/中文文献进行梳理,总结MTO研究进展和趋势;然后,基于多因子优化(multifactorial optimization,MFO)和多种群演化(multi-population evolution,MPE)两种不同信息共享框架,从多任务搜索空间设计、种群数量、种群规模、依托算法、信息迁移节点、交互信息、时间和空间复杂度以及复杂系统等角度对比二者异同;接着,从信息迁移节点、方式和类型3方面重点阐述MTO核心理论;最后,从探究MTO复杂系统层级智能涌现行为、多任务种群多样性控制以及应用领域拓展3方面展望未来研究方向.
Multi-task optimization(MTO)based on implicit parallelism of computational intelligence is a research hotspot and cutting-edge technology nowadays.Compared with the traditional single-task optimization algorithm,optimizing multiple tasks concurrently by mining the internal parallelism and connotation of swarm intelligence can significantly improve the problem solving quality and shorten the task solving time.Firstly,the research trend and progress of the MTO is outlined by combing the English and Chinese related literature.Then,based on the multifactorial optimization(MFO)and multi-population evolution(MPE)information sharing frameworks,the differences and similarities between the MFO and the MPE are summarized from the aspects of multi-task search space design,population number,population size,relying algorithms,information migration nodes,transferred knowledge,time and spatial complexity,complex systems.Moreover,the core theory of MTO is expounded from the angel of information transfer nodes,modes and types.Finally,the future research direction from the aspects of exploring the hierarchical intelligent emergence of the MTO complex system,multi-task population diversity control and expanding application fields are discussed.
作者
程美英
钱乾
倪志伟
CHENG Mei-ying;QIAN Qian;NI Zhi-wei(School of Economics&Management,Huzhou University,Huzhou 313000,China;School of Teacher Education,Huzhou University,Huzhou 313000,China;School of Management,Hefei University of Technology,Hefei 230009,China)
出处
《控制与决策》
EI
CSCD
北大核心
2023年第7期1802-1815,共14页
Control and Decision
基金
国家自然科学基金青年基金项目(62102148)
湖州市科技计划项目(2018YZ11).
关键词
多任务优化
信息共享框架
信息迁移节点
信息迁移方式
信息迁移类型
典型应用领域
multi-task optimization
information sharing framework
information transfer node
information transfer mode
information transfer type
typical application area
