流水车间调度问题的快速多目标混合进化算法

针对最大完工时间最小和总流经时间最小的双目标流水车间调度问题,提出一种快速多目标混合进化算法。算法将矢量评价遗传算法的采样策略与一种新的基于Pareto支配与被支配关系的适应度函数的采样策略进行了融合。新的采样策略弥补了矢量评价遗传算法(VEGA)采样策略的不足。VEGA善于搜索Pareto前沿面的边缘区域,但却忽略了Pareto前沿面的中心区域,而新的采样策略则倾向于Pareto前沿面的中心区域。这两种机制的融合保证了混合算法能够快速平稳地向Pareto前沿区域收敛。此外,由于混合采样策略不需要考虑距离,使得算法效率也得到了很大的提升。在对Taillard基准测试集进行的仿真实验结果显示,相对于非支配排序遗传算法(NSGA-Ⅱ)和强度Pareto进化算法(SPEA2),该快速多目标混合进化算法在收敛性和分布性两方面都有所提高,并且算法的效率也得到了改进。所提出的混合算法能够更好地解决双目标的流水车间调度问题。     

基于流固耦合与多目标拓扑优化的低噪声塑料机油冷却器盖优化设计

机油冷却器盖属于薄壁件,距离振动激励源较近,容易产生较大的振动噪声,且内腔冷却液的存在对机油冷却器盖的振动噪声有着很大的影响。为有效的对塑料机油冷却器盖的振动噪声进行仿真研究及优化,将流体冲击压力作为预应力的同时,结合塑料机油冷却器盖与内腔流体的流固耦合模型,采用流固耦合的方法对塑料机油冷却器盖的振动噪声水平进行了预测评估;根据预测结果,识别出对噪声贡献度较大的耦合模态频率;以降低塑料机油冷却器盖整体噪声为总目标,以提高各贡献度较大的耦合模态频率为子目标,利用加权指数法建立了多目标优化函数,对塑料机油冷却器盖结构进行了多目标拓扑优化。结果表明,优化后降噪效果明显,总声功率级降低了1.79 dB。     

基于多目标遗传算法的炼焦生产过程优化控制

针对某钢铁企业实际炼焦生产过程的优化控制问题,提出一种多目标分层优化控制策略.首先采用主成分分析、灰色关联分析及改进前馈神经网络方法,建立综合生产目标与局部优化目标的关联模型,将综合生产目标映射为局部优化目标.然后建立以焦炭产量最大、焦炉能耗最小为优化目标,焦炭质量与工艺要求为约束条件,局部优化目标为决策变量的多目标优化模型.通过多目标遗传算法求解多目标优化问题,获得局部优化目标值.最后将局部优化目标作为各子过程控制系统的设定值及优化调度系统的决策参数,来动态调整过程操作参数,实现企业期望的综合生产目标.实际运行结果表明,提出的优化控制策略取得了较好的应用效果.     

智能型电网事故辅助决策系统在调度中的应用

原有的电力调度自动化系统已不能适应当前电力系统发展和电力体制改革的需要。介绍了智能型电网事故辅助决策系统的提出背景、组成模块和实现的功能。通过在线模拟案例,阐述了该决策系统对实现智能调度的意义。     

车门板辊压包边机器人环境力约束仿真与试验研究

介绍了采用机器人对车门板进行辊压包边的生产工艺。基于有限元方法对辊压包边过程中板料的变形抗力进行了计算,对辊轮三种可能的压入方式（竖直、水平、环向）进行仿真分析,得出了在不同压入方式下包边机器人末端所受的最大环境约束力,同时在万能材料试验机上进行了相应试验。计算和试验结果表明,竖直压入方式辊轮所受环境力峰值最大,环向压入方式环境力峰值最小。     

Aggregate production planning for shipbuilding with variation-inventory trade-offs

Shipbuilding is a complex production system characterised by a complicated work and organisation structure, prolonged production lead time, and heterogeneous resource requirements. Thus, effectively planning all involved activities presents a challenging task and requires the timely coordination between the successive production stages at the plant level and effective resource allocation at the workshop level. With the work breakdown structure of all projects and their corresponding building strategies, the aggregate production planning (APP) is to address two important issues, namely, workforce level and inventory usage so that the fluctuating demands from downstream processes can be satisfied in a cost-effective manner. To achieve this, a novel APP model is proposed for ship production to minimise the variation of aggregate man-hour over the planning horizon and simultaneously minimise the logistic demands of the interim products. In view of the combinatorial nature and computational complexity, a directed genetic algorithm based solver has been developed to solve the two-conflicting-objective optimisation problem. The proposed approach has been applied to a case study and preliminary results have shown certain effectiveness in handling various situations with different planning strategies.     

A hybrid Pareto-based local search algorithm for multi-objective flexible job shop scheduling problems

This paper presents a hybrid Pareto-based local search (PLS) algorithm for solving the multi-objective flexible job shop scheduling problem. Three minimisation objectives are considered simultaneously, i.e. the maximum completion time (makespan), the total workload of all machines, and the workload of the critical machine. In this study, several well-designed neighbouring approaches are proposed, which consider the problem characteristics and thus can hold fast convergence ability while keep the population with a certain level of quality and diversity. Moreover, a variable neighbourhood search (VNS) based self-adaptive strategy is embedded in the hybrid algorithm to utilise the neighbouring approaches efficiently. Then, an external Pareto archive is developed to record the non-dominated solutions found so far. In addition, a speed-up method is devised to update the Pareto archive set. Experimental results on several well-known benchmarks show the efficiency of the proposed hybrid algorithm. It is concluded that the PLS algorithm is superior to the very recent algorithms, in term of both search quality and computational efficiency.     

Assembly line balancing by a new multi-objective differential evolution algorithm based on TOPSIS

In this paper, we propose a multi-objective differential evolution algorithm (MODEA) to solve the multi-objective simple assembly line balancing problem type-2 (SALBP-2). This problem arises when in an existing assembly line, changes in the production process or demand structure take place and the organisation wants to produce the optimum number of items using a fixed number of workstations, which is associated with optimally assigning the tasks to an ordered sequence of stations such that the precedence relations are not violated and some measures of performance are optimised. The two considered objectives are: minimising the cycle time and the smoothness index of the assembly line. To that purpose, we develop a MODEA which unlike the existing algorithms deals with the considered objectives separately in selecting the next population members by proposing a new acceptance scheme based on the Pareto dominance concept and a new evaluation scheme based on TOPSIS. Also, by using the Taguchi method, we tune the effective factors of the developed algorithm. Then its efficiency is tested over available assembly line balancing benchmarks and compared to a new algorithm provided recently in the bi-objective SALBP-2 literature. Computational experiments indicate that the developed algorithm outperforms the existing meta-heuristic over a large group of benchmarks.     

Collaborative capacity sharing among manufacturers on the same supply network horizontal layer for sustainable and balanced returns

To deal with volatile demand and rapidly changing manufacturing technologies for sustainable returns, as a solution, collaborative capacity sharing (CCS) among manufacturers at the same horizontal layer in supply networks is discussed in this research, especially focusing on a long-term aspect. Such horizontal collaboration can minimise manufacturers’ lost sales, as well as maximise their production capacity utilisation in a long-term period against lumpy demand. Hence, to help manufacturers determine whether and with whom they have to collaborate a CCS protocol based on a theoretical analysis is developed. It is a distributed decision protocol suggesting an effective coalition for each manufacturer. Furthermore, for well-balanced distribution of resulting profits to each manufacturer, two procedures are proposed: (1) determination of appropriate substitution production cost and (2) bids among manufacturers/coalitions. To evaluate the effect of those processes, two types of CCS protocols are designed – CCS-(1) and CCS-(2); only CCS-(2) performs those procedures. A numerical experiment is conducted to compare the performance of four models: No collaboration, complete collaboration, and two types of selective collaboration by CCS-(1) and CCS-(2). The performance of CCS-(2) is at least equal to or better than other models with higher and balanced returns.     

Inventory-shortage driven optimisation for product configuration variation

Inventory control is a critical problem in manufacturing systems. Inventory shortage significantly affects system productivity, while excessive stocks increase the operation cost. It is difficult to avoid fully inventory shortage under mass customisation manufacturing based on product configuration. In this paper, we propose a new approach for inventory-shortage driven optimisation of dynamic product configuration variation to meet the requirements of product configuration change and find suitable combination of parts by considering cost, lead-time and inventory variation. The multi-objective optimisation model uses a multi-objective genetic algorithm and adds impact cost, lead-time and inventory factors to the normal configuration optimisation model. An industrial case study demonstrates the practicality and effectiveness of the proposed approach. By means of this research, valid solutions for configuration variation are available to the decision makers.