求解物流配送同时取送货低碳选址—路径问题的量子超启发式算法  

Quantum-inspired hyper-heuristics algorithm for low-carbon location-routing problem with simultaneous pickup and delivery

作　　者：冷龙龙 赵燕伟[1] 蒋海青 张春苗 王舜[1] LENG Longlong;ZHAO Yanwei;JIANG Haiqing;ZHANG Chunmiao;WANG Shun(College of Mechanical Engineering,Zhejiang University of Technology,Hangzhou 310014,China;Mechanical and Automotive Branches,Jiaxing Vocational and Technical College,Jiaxing 314036,China)

机构地区：[1]浙江工业大学机械工程学院,浙江杭州310014 [2]嘉兴职业技术学院机电与汽车分院,浙江嘉兴314036

出　　处：《计算机集成制造系统》2020年第3期815-828,共14页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(61572438);浙江省科技计划资助项目(2017C33224)。

摘　　要：为降低物流配送过程中车辆排放量,建立以碳排放为目标的同时取送货低碳选址—路径问题模型,并利用量子超启发式算法对模型进行求解。量子超启发算法以量子进化策略作为超启发式算法的高层学习策略,并结合滑动窗口机制实现底层算子的准确搜索,以提高算法框架性能。此外,为了减少计算负担,构造了能够保证可行性解的编码方式和底层算子,并不需要任何解的修复技术,在此基础上提出一种快速简单易行的适应度评价方法。通过不同规模的实例的仿真实验和对比分析,表明所提算法能够在合理的时间内获得优质解,并验证了该算法的有效性和鲁棒性。Aiming at reducing the carbon emission of vehicles,the Location-Routing Problem with Simultaneous Pickup and Delivery(LRPSPD)model by taking minimum carbon emission as objective was defined,which was tackled by Quantum-inspired Hyper-Heuristic(QHH).In framework of QHH,the quantum Q-gate mechanism based sliding windows was developed as the hyper-level learning strategy to timely manipulate and track the recent performance of Low-Level Heuristics(LLH)for pinpointing the appropriate operators,which was utilized to implement a sort of simple operators such as‘shift’and‘swap’.Moreover,for reducing the computational burden,a solution representation and LLHs were designed to guarantee feasibility of solutions and avoid the use of repair methods to restore feasibility,and a fast solution evaluation technique was developed.Simulation results and comparisons showed that the proposed algorithms were effective,and provided promising results for different scales instances within reasonable computing time.

关 键 词：选址—路径问题 同时取送货 低碳物流 超启发式算法 量子选择策略 

分 类 号：F224[经济管理—国民经济] X24[环境科学与工程—环境科学]