钢铁一体化生产调度模型与优化算法北大核心CSSCI

为达到降低在制品库存及保证生产连续性等目标,将炼钢-连铸-热轧一体化生产过程归结为中间阶段为批处理机的多阶段柔性流水车间调度问题,建立了一体化生产模型。该模型以总加权完成时间和总加权滞留时间之和为优化目标,综合考虑了机器加工能力、浇次处理要求、炉次释放时间等约束条件。将拉格朗日松弛算法应用于模型中,利用该算法的批分解方式和子问题的双向动态规划求解策略,提高了算法的收敛速度和下界。不同规模问题的测试结果表明了模型和算法的有效性。     

钢铁一体化生产调度模型与优化算法

为达到降低在制品库存及保证生产连续性等目标,将炼钢-连铸-热轧一体化生产过程归结为中间阶段为批处理机的多阶段柔性流水车间调度问题,建立了一体化生产模型。该模型以总加权完成时间和总加权滞留时间之和为优化目标,综合考虑了机器加工能力、浇次处理要求、炉次释放时间等约束条件。将拉格朗日松弛算法应用于模型中,利用该算法的批分解方式和子问题的双向动态规划求解策略,提高了算法的收敛速度和下界。不同规模问题的测试结果表明了模型和算法的有效性。     

考虑新任务插入的炼钢-连铸重调度模型与算法

针对新任务插入的炼钢-连铸重调度问题,将新任务插入映射为增加的时间约束,以开工时间差异化最小为目标建立了动态约束满足模型,提出了基于动态约束满足的局部修复算法。算法在变量选择中引入最小自由度优先策略用于选择变量,在值选择中提出最大空闲时间量可行分配策略用于给变量赋值,利用冲突识别规则识别赋值过程中产生的时序冲突和资源冲突并用冲突解消规则解消冲突。仿真实验模拟了3个具有代表性的插入时刻点,所得目标函数值均为零。结果表明所提算法能够最大限度地满足满足生产的实时性和稳定性的需求。     

考虑工序并行的差异工件批调度研究

针对铸造车间差异工件组批多约束的问题,在工序可并行加工的前提下构建以最小化最大完工时间和最小化沙箱空置率为优化目标的并行工序批调度模型,设计一种改进和声算法求解该调度模型,提出一种单工序编解码方式和2种机器分配规则用于解决工件分批、沙箱选择、工序分配及机器选择的问题。在算法中提出一种新的和声产生方式和更新机制,同时为改善算法的局部搜索能力,加入模拟退火算法执行局部搜索过程。最后根据企业实际生产数据进行仿真实验,验证本文模型的有效性。     

不确定环境下等待时间受限的混合流水车间调度问题研究

带有等待时间受限约束的混合流水车间调度问题具有强NP难的特性,而实际生产过程中产品需求时间以及机器加工时间等各种不确定性因素,更为求解此类问题带来了复杂性和挑战。通过探索这类问题的特征,建立了两阶段随机规划模型,开发了L型切面的求解算法,在目标为成本期望最小的情况下,给出了这类问题的近似最优调度解。建立的模型针对于生产过程中的稳态,以更好地适应实际生产的需要。通过对算法性能进行理论分析和实验验证,证明了该算法的有效性。     

生产批量计划问题的干扰管理研究

以成组单元生产批量计划问题为原型,研究了在生产计划执行过程中受到机器故障干扰的调整方案.运用干扰管理方法,在扰动判定和度量的基础上,建立了考虑计划偏离度和总成本最小化的成组单元生产批量计划干扰管理模型,并设计了求解该模型的知识进化算法.当干扰发生时,系统能快速响应因机器故障造成的生产能力不足问题,重新调整生产计划.通过仿真实验,与全局重调度方案结果进行对比,验证了本干扰事件处理方法的可行性和有效性.     

钢铁企业订单动态交货期承诺模型与算法

针对钢铁企业交货期承诺问题的动态特征,建立了动态约束满足模型.该模型综合考虑了设备利用率和订单的提前拖期惩罚.模型求解属于NP-hard问题,很难用精确算法在可行时间内求解的特点,因此结合启发式规则和约束满足一致性技术的优点,提出了求解该问题的算法.通过实验验证了模型及算法的可行性和有效性.     

面向汽车整车出厂物流的两阶段组合装箱调度优化模型与算法研究

研究组合装箱与道位分配的两阶段出厂物流的调度模型与算法,问题将考虑订单、驳运车、仓库道位资源、客户需求、配送成本等多种因素。首先将所研究问题划分为带装箱约束、方向约束和拼车约束的组合装箱子问题,和带驳运车到达时间及道位不可用时间段的道位分配问题。针对两个子问题建混合线性整数模型并分别设计了分支定界和启发式算法,实现两阶段整合求解。最后,通过数值实验和与人工调度结果的对比模型,表明了模型和算法的有效性。     

集装箱码头岸桥调度的时间—成本均衡模型与算法

针对集装箱码头岸桥调度的时间—成本均衡问题,综合考虑岸桥调度问题中所涉及的各类约束和装卸任务所需时间与成本之间的关系,首次提出了岸桥调度的时间—成本均衡混合整数优化模型。设计改进的遗传算法用以求解该模型。数值实验的结果验证了模型正确性和算法有效性,说明模型和算法可以用于实际的岸桥调度过程。     

循环取货带有时间窗约束的入库道口车辆调度

研究了循环取货模式下带有时间窗约束的入库道口车辆调度问题,为使车辆运输成本和取货时间成本、卸货时间成本最小,建立混合整数规划数学模型,设计了两阶段算法求解模型,第一阶段产生满足容量约束的较好初始解,第二阶段通过发车时间与路径同时编码的模拟退火算法进行求解,根据某汽车制造商循环取货的实际运作情况,构造算例并验证了该模型和算法的有效性。结论表明,制造商处的道口限制对循环取货发车时间与路径调度有较大影响,同时对发车时间和路径进行调度更有利于降低循环取货的运输费用。     

Dynamic scheduling in steelmaking-continuous casting production for continuous caster breakdown

In the steelmaking-continuous casting (SCC) production process, machine breakdown is one of the most common disturbances which may make the current schedule unrealisable. Existing rescheduling models for machine breakdown only employ one constraint that charges cannot be processed on this machine in its failure period. However, this method is effective for steelmaking furnace breakdown and refining furnace breakdown but invalid for continuous caster breakdown. Due to the production characteristics of continuous caster, reallocating a casting order and a continuous caster for each unfinished charge on the broken down continuous caster is necessary before making a new schedule. Different reallocation strategies have different impacts on charge’s processing time and processing stage route in the dynamic scheduling process. Therefore, SCC dynamic scheduling for the continuous caster breakdown is different from the other machines. In this paper, the impacts of these strategies are studied, and a dynamic scheduling model which can be used to generate a new schedule for each strategy is built. To obtain a high-quality solution in acceptable computational time for this model with NP-hard feature, a hybrid algorithm featuring a genetic algorithm combined with a general variable neighbourhood search is developed based on the problem-specific characteristics. Computational experiments on practical production data show that the proposed rescheduling method is effective for SCC dynamic scheduling with continuous caster breakdown.     

反相容工件系统的加权完工时间和的重新排序问题(英文)

重新排序问题是指在原始工件已经安排好的情形下,新到的工件集与原始工件集一起重新再排序,这是实际工作中常见一类优化问题。本文考虑了单机上当工件加工时间与权重反相容时,在最大错位量约束下的加权完工时间和最小化的重新排序问题。对于提出的四个问题,即在最大序列错位、最大时间错位、总序列错位和总时间错位约束下的加权完工时间和重新排序,基于问题的结构性质,运用动态规划方法分别给出了这些问题的多项式时间或拟多项式时间算法。     

Application of a Solidification Mathematical Model and a Genetic Algorithm in the Optimization of Strand Thermal Profile Along the Continuous Casting of Steel

This work presents an optimization method based on a genetic algorithm applied to continuous casting process. A simple genetic algorithm was developed, which works linked to a mathematical model permitting the determination of optimum values for the water flow rates in the secondary cooling zones. First, experimental data (industrial) were compared with simulated results obtained by the solidification mathematical model, to determine the metal/cooling heat transfer coefficients along the machine by the inverse heat conduction problem method. The industrial data concerning surface strand temperature were obtained by using infrared pyrometers along a continuous caster machine during casting of both SAE 1007 and 1025 steels. In a second step, these results were used by a numerical code based on a genetic algorithm for determining optimum settings of water flow rates in the different sprays zones, which are conducive to the best quality of the solidified strand. The simulations were carried out by analyzing the solidification process during continuous casting to attain metallurgical restrictions concerning the reheating of strand surface temperature and metallurgical length.     

Rescheduling with controllable processing times for number of disrupted jobs and manufacturing cost objectives

We consider a machine rescheduling problem that arises when a disruption such as machine breakdown occurs to a given schedule. Machine unavailability due to a breakdown requires repairing the schedule as the original schedule becomes infeasible. When repairing a disrupted schedule a desirable goal is to complete each disrupted job on time, i.e. not later than the planned completion time in the original schedule. We consider the case where processing times of jobs are controllable and compressing the processing time of a job requires extra processing cost. Usually, there exists a nonlinear relation between the processing time and manufacturing cost. We solve a bicriteria rescheduling problem that trades off the number of on-time jobs and manufacturing cost objectives. We give a mixed-integer second-order cone programming formulation for the problem. We develop a heuristic search algorithm to generate efficient solutions for the problem. Heuristic algorithm searches solution space by moving and swapping jobs among machines. We develop cost change estimates for job moves and swaps so that the heuristic implements only promising moves and hence generates a set of efficient solutions in reasonably short CPU times.     

Research on rush order insertion rescheduling problem under hybrid flow shop based on NSGA-III

Rush order insertion is widespread in the enterprises that apply make-to-order production mode which affects the stability of production system. This article studies rush order insertion rescheduling problem (ROIRP) under hybrid flow shop (HFS) with multiple stages and multiple machines. A mathematical model simultaneously considering constraints such as lots, sequence-dependent set-up times and transportation times with objectives to minimise makespan, total transportation time and total machine deviation between the initial scheduling plan and the event-driven rescheduling plan is developed and NSGA-III is applied to solve the problem. Three groups of experiments are carried out which verify the suitability of NSGA-III for HFS scheduling problem with multi-objective and multi-constraint, the effectiveness of NSGA-III for the proposed ROIRP and the feasibility and effectiveness of the proposed model and algorithm in solving the ROIRP of a realistic ship pipe parts manufacturing enterprise.     

Models and optimisation approaches for scheduling steelmaking–refining–continuous casting production under variable electricity price

This paper studies the steelmaking–refining–continuous casting (SRCC) scheduling problem with considering variable electricity price (SRCCSPVEP). SRCC is one of the critical production processes for steel manufacturing and energy intensive. Combining the technical rules used in iron-steel production practice, time-dependent electricity price is considered to reduce the electricity cost and the associate production cost. A decomposition approach is proposed for the SRCCSPVEP. Without considering the electrical factor, the first phase applies the mathematical programming method to determine the relative schedule plan for each cast. In the second phase, we formulate a scheduling problem of all casts subject to resource constraint and time-dependent electricity price. A heuristic algorithm combined with the constraint propagation is developed to solve this scheduling problem. To investigate and measure the performance of the proposed approach, numerous instances are randomly generated according to the collective data from a well-known iron-steel plant in China. Computational results demonstrate that our algorithm is very efficient and effective in providing high-quality scheduling plans, and the electricity cost can be reduced for the iron-steel plant.     

Parallel machine scheduling with limited controllable machine availability

Most machine scheduling models assume that either the machines are available all the time, or the time of their unavailability is fixed as a constraint. In this paper, we study the problem that neither the unavailability length nor the start time of machine unavailability is fixed. Instead, they would be determined in order to minimise the total cost involved with the completion time and the unavailable time. This model could represent a more realistic and complex situation, in which jobs and machines’ availability operations should be optimised simultaneously. After the model is formulated, some properties of the problem are presented. Then a branch and bound algorithm based on column generation approach is proposed to solve the problem. The computation results show that, within a reasonable computation time, the proposed algorithm can solve medium sized problems optimally.     

A differential evolution algorithm for the no-idle flowshop scheduling problem with total tardiness criterion

In this paper, we investigate the use of a continuous algorithm for the no-idle permutation flowshop scheduling (NIPFS) problem with tardiness criterion. For this purpose, a differential evolution algorithm with variable parameter search (vpsDE) is developed to be compared to a well-known random key genetic algorithm (RKGA) from the literature. The motivation is due to the fact that a continuous DE can be very competitive for the problems where RKGAs are well suited. As an application area, we choose the NIPFS problem with the total tardiness criterion in which there is no literature on it to the best of our knowledge. The NIPFS problem is a variant of the well-known permutation flowshop (PFSP) scheduling problem where idle time is not allowed on machines. In other words, the start time of processing the first job on a given machine must be delayed in order to satisfy the no-idle constraint. The paper presents the following contributions. First of all, a continuous optimisation algorithm is used to solve a combinatorial optimisation problem where some efficient methods of converting a continuous vector to a discrete job permutation and vice versa are presented. These methods are not problem specific and can be employed in any continuous algorithm to tackle the permutation type of optimisation problems. Secondly, a variable parameter search is introduced for the differential evolution algorithm which significantly accelerates the search process for global optimisation and enhances the solution quality. Thirdly, some novel ways of calculating the total tardiness from makespan are introduced for the NIPFS problem. The performance of vpsDE is evaluated against a well-known RKGA from the literature. The computational results show its highly competitive performance when compared to RKGA. It is shown in this paper that the vpsDE performs better than the RKGA, thus providing an alternative solution approach to the literature that the RKGA can be well suited.     

Auction-based approach to resolve the scheduling problem in the steel making process

Steel production is an extremely complex process and determining coherent schedules for the wide variety of production steps in a dynamic environment, where disturbances frequently occur, is a challenging task. In the steel production process, the blast furnace continuously produces liquid iron, which is transformed into liquid steel in the melt shop. The majority of the molten steel passes through a continuous caster to form large steel slabs, which are rolled into coils in the hot strip mill. The scheduling system of these processes has very different objectives and constraints, and operates in an environment where there is a substantial quantity of real-time information concerning production failures and customer requests. The steel making process, which includes steel making followed by continuous casting, is generally the main bottleneck in steel production. Therefore, comprehensive scheduling of this process is critical to improve the quality and productivity of the entire production system. This paper addresses the scheduling problem in the steel making process. The methodology of winner determination using the combinatorial auction process is employed to solve the aforementioned problem. In the combinatorial auction, allowing bidding on a combination of assets offers a way of enhancing the efficiency of allocating the assets. In this paper, the scheduling problem in steel making has been formulated as a linear integer program to determine the scheduling sequence for different charges. Bids are then obtained for sequencing the charges. Next, a heuristic approach is used to evaluate the bids. The computational results show that our algorithm can obtain optimal or near-optimal solutions for combinatorial problems in a reasonable computation time. The proposed algorithm has been verified by a case study.