人机协同双边装配的能力规划与作业调度

In recent years, many automobile manufacturers in China have set up advanced man-robot cooperative two-sided assembly systems. However, it is difficult to obtain efficient assembly schedule by relying on the worker experience-based heuristic. This might lead to many problems such as the idleness of man and robot and production line interruption..Aiming at the man-robot cooperative two-sided assembly mixed production system, this research explores the intelligent scheduling method with the goal of optimization efficiency, cost and stability. This study is carried out mainly from the following three levels. (1) Firstly, this project investigates man-robot assembly capabilities and system-specific constraints, establishes the stability indexes in the man-robotic cooperative two-sided assembly line, and builds the multi-objective integrated system based on mixed integer programming method. (2) Secondly, this project classifies and analyzes uncertain disturbances, designs preventive and reactive scheduling strategies, and explores adaptive scheduling strategy that combines robust scheduling, partial repair or remediation, and full re-scheduling. (3) Thirdly, this project explores the skeleton and fitness landscape of the considered problem, constructs the multi-constrained multi-objective intelligent scheduling method with coupling problem characteristics and swarm intelligence, and analyzes the algorithm complexity, convergence and other performance. The research results will enrich the existing assembly scheduling theory, and it is expected to enhance the production management level of China's automobile assembly and promote the automation and intelligence of the assembly manufacturing industry in our country.

近年来，国内多家汽车生产企业组建了先进的人机协同双边装配系统，但是现有依靠工人经验的调度方式难以获得高效的装配方案，易造成装配系统人机闲置、产线中断等问题。.本项目面向人机协同双边混装生产系统，探究以效率、成本和平稳为目标的智能调度方法。主要从三个层面实现：①挖掘人机装配能力特性、系统特有约束，建立人机协同双边装配系统平稳指标体系，构建基于混合整数规划的多目标集成系统模型；②归类和分析不确定扰动，设计预防调度及反应调度策略，探索融合鲁棒调度、局部修复或补救、完全重调度的自适应调度策略；③探索问题的骨架和适应度地貌等特征，构建耦合问题特征和群智能算法特性的多约束多目标智能调度算法，并进行算法复杂度、收敛性等性能分析。研究成果将丰富现有的装配调度理论，有望提升我国汽车装配的生产管理水平，促进我国装配制造业的自动化和智能化。

国内多家汽车生产企业组建了先进的人机协同双边装配系统，但是现有依靠工人经验的调度方式难以获得高效的装配方案，易造成装配系统人机闲置、产线中断等问题。围绕着人机协同双边装配线，1) 项目组首先挖掘人机装配能力特性、系统特有约束等，按照逐层深入的方法，依次构建了机器人装配线、多技能员工装配线和人机协作装配线集成优化模型，并采用CPLEX求解器实现小规模案例的精确求解以验证模型的正确性与通用性；2) 其次，针对现场的不确定扰动，设计了考虑预防维护的装配线平衡优化技术，以提升整条混流装配线的效率和稳定性，同时设计了考虑装配线中断的装配线再平衡技术，以便在装配线出现停机故障时选择并快速获得新的调度方案；3) 最后，项目组研究了多约束处理机制、算法算子设计、算法优化机理、多目标处理机制，进而设计了高效的精确算法、单目标局部搜索算法、单目标群智能算法和多目标群智能算法，以便快速可靠地生成调度方案，更好地满足企业需求。.特别的，项目组在专著中对人机协作双边装配线的模型和方法进行了全面归纳，其中人机协作具体被划分为四种情形：1) 情形一：共享工位、独立协同和不可并行执行操作；2) 情形二： 共享工位、独立协同和可并行执行操作；3) 情形三：共享工位、可合作协同和不可并行执行操作；4) 情形四：共享工位、可合作协同和可并行执行操作。针对每种情形，项目组构建了问题的集成调度模型，同时基于问题特征分别构建了单目标单解智能算法、单目标群智能算法和多目标群智能算法。.围绕着人机协作装配线平衡问题，项目组在《European Journal of Operational Research》、《Journal of Intelligent Manufacturing》、《Expert Systems With Applications》、《Neural Computing and Applications》、《Applied Soft Computing Journal》及《Applied Mathematical Modelling》等重要刊物发表SCI论文17篇，发表专著1本。合作开发了名为Line BalanceR的装配线平衡优化商业软件。培养博士生2名，同时两名硕士生转博。