Allocating kanbans for a production system in a general configuration with a new control strategy

We consider a production system in a general configuration with a new control strategy: the push policy for the part sending and the kanban mechanism for the work-in-process (WIP). The production system is composed of many stations (or workshops) such as an entry station, a set of workstations, a central station, and an exit station. This type of system is modeled as an open queueing network (OQN) in a general configuration with a Markov-type part sending policy and a machine no blocking (MNB) mechanism. The most important performance measures of the production system are the total throughput of the workstations and the total blocking flow of blocked parts sent from the workstations to the central station. This paper discusses an optimization problem with multiple objectives: allocate kanbans to the workstations so as to simultaneously maximize the total throughput and minimize the total blocking flow. Based on a semi-open decomposition approach, several useful properties of the system are characterized. These properties are used to develop a marginal algorithm for the optimization problem. Moreover, a dynamic simulation approach is devised as a tool for evaluating the quality of the solutions obtained by the algorithm. Numerical experiments are provided to demonstrate the efficiency of the algorithm through the simulation approach.     

The optimal location of inspection stations using a rule-based methodology

The location of inspection stations is a significant component of production systems. In this paper, a prototype expert system is designed for deciding the optimal location of inspection stations. The production system is defined as a single channel of n serial operation stations. The potential inspection station can be located after any of the operation stations. Non-conforming units are generated from a compound binomial distribution with known parameters at any given operation station.

Traditionally Dynamic programming, Zero-one integer programming or Non-linear programming techniques are used to solve this problem. However a problem using these techniques is that the computation time becomes prohibitively large when the number of potential inspection stations are fifteen or more. An expert system has the potential to solve this problem using a rule-based system to determine the near optimal location of inspection stations.

The prototype expert system is divided into a static database, dynamic database and knowledge base. Based on defined production systems, the sophisticated rules are generated by the simulator as part of a knowledge base. A generate-and-test inference mechanism is utilized to search the solution space by applying appropriate symbolic and quantitative rules. The goal of the system is to determine the location of inspection stations while minimizing total cost.     

Buffer sizing of a Heijunka Kanban system

Heijunka is a key-element of the Toyota production system which levels the release of production kanbans in order to achieve an even production flow over all possible types of products, thus, e.g. reducing the bullwhip effect. In this paper we analyze a kanban controlled and heijunka leveled production system where the arriving demands are controlled and limited by a kanban loop. The production system is modeled as a queueing network with synchronization stations. The aim is to determine the optimal number of production kanbans, and thus the buffer size that guarantees a given service level.     

A Simulated Annealing algorithm for a mixed model assembly U-line balancing type-I problem considering human efficiency and Just-In-Time approach

This research deals with balancing a mixed-model U-line in a Just-In-Time (JIT) production system. The research intends to reduce the number of stations via balancing the workload and maximizing the weighted efficiency, which both are considered as the objectives of this research paper.After balancing the line and determining the number of stations, the labor assignment policy should be set. In this study, it was assumed that there are two types of operators: permanent and temporary. Both types can work in regular and overtime periods. Based on their skill levels, workers are classified into four types. The sign at each work station indicates types of workers allowed to work at that station. An alert system using the hybrid kanban systems was also considered. To solve this problem, a Simulated Annealing algorithm was applied in the following three stages. First, the balancing problem was solved and the number of stations was determined. Second, workers were assigned to the workstations in which they are qualified to work. Following that, an alert system based on the kanban system was designed to balance the work in the process inventory. This was achieved by defining control points based on the processing time and making control decisions to minimize the number of kanban cards. In the proposed SA algorithm, two methods for the temperature cooling schedule were considered and two methods were defined for determining the number of neighborhood search. The initial temperature was considered equal to the cost of the initial solution to reach the convergence situation as soon as possible. Five problems were solved in small size using the GAMS software. The results obtained from the GAMS software were compared with those obtained from the SA algorithm to determine the performance difference. The computational results demonstrated that the SA algorithm is more consistent with the answers obtained. Also seven large scale problems were solved. The results showed that the SA algorithm still have better reliability. To show the efficiency of the proposed SA algorithm, an axel assembly company was studied. To satisfy demands and reduce backlogging, a mixed model assembly line was designed for this case study. The results showed that the mixed model assembly line designed using the SA algorithm had good efficiency.     

Algorithm for the design of single-stage adaptive kanban system

The traditional kanban system with fixed number of cards does not work satisfactorily in unstable environment. In the adaptive kanban-type pull control mechanism the number of kanban is allowed to change with respect to the inventory and backorder level. It is required to set the threshold values at which cards are added or deleted which is a part of the design. Previous studies used the local search method to design the adaptive kanban system. In this paper Genetic Algorithm- and Simulated annealing-based heuristics are developed and used to set the design parameters of adaptive kanban system. The numerical results indicate that simulated annealing based heuristics produces better solution with improved computational efficiency.     

基于性能分析的装配站生产状态快速配置方法

为了支持装配生产作业中市场需求、生产率和资源利用率的协调控制,保证企业的高效、精益生产,提出了一种基于性能分析的装配站生产状态快速配置方法。该方法首先构建支持性能分析的装配站作业网络图,实现在制品资源与可重用资源双重约束下的作业性能分析;然后,考虑闲置时间和成本浪费两类指标,给出装配站最优生产状态评价的多目标函数,并探讨问题求解的解空间特性;采用在制品控制和随机领域两个改进策略,建立生产状态快速配置的粒子群算法。最后,采用某型飞机机翼装配站为实例,验证了方法的有效性。     

On the workload and ‘phaseload’ allocation problems of short reliable production lines with finite buffers

Serial flow or production lines are modeled as tandem queueing networks and formulated as continuous-time Markov chains to investigate how to maximize throughput or minimize the average work-in-process (WIP) when the total service time and the total number of service phases among the stations are fixed (these are the workload and ‘phaseload’ allocation problems, respectively). This paper examines both the effect of the kind of service time distribution on the optimal workload allocation in order to maximize throughput or minimize the average WIP of perfectly reliable production lines.

The new approach of this work is the differentiation of the number of service phases of the service time distribution which is assumed to be of phase type at all stations of the flow line and the placement of storage space (buffers) between any two successive stations in order to examine the effect of these factors to the form of the optimal workload vector t and the optimal service phase vector ph.     

电动汽车典型快充站优化运行配置方法

为了降低大功率快充桩对电网冲击波动,并考虑典型快充站分布式电源和储能优势,提出一种电动汽车典型快充站优化运行配置方法.通过分析站内分布式电源出力特点以及电动汽车充电行为规律,以充电站运行成本最小为优化目标,建立典型快充站优化运行配置模型,以站内功率平衡、分布式电源出力等为约束条件,利用遗传优化算法求解模型最优解.最后通过不同配置算例验证所提方法的可行性,为典型快充站的优化运行提供技术支撑.     

The determination of the optimal number of kanbans in a Just-In-Time production system

Just-In-Time (JIT) production is supported by the kanban inventory control and product scheduling system. It is important for production managers to determine the optimal number of kanbans in order to successfully operate a multi-line, multi-stage kanban production system. All relevant factors and costs must be considered before arriving at the optimal number of kanbans. In this paper, some JIT production factors such as demand, in process inventory, inventory and labor costs, subcontractor's supply capacity, and workload are discussed. And then, a model formulation using mixed integer goal programming is presented. Finally, a numerical example is presented in order to verify our model.     

Arrival first queueing networks with applications in kanban production systems

In this paper, we introduce a new class of queueing networks called arrival first networks. We characterise its transition rates and derive the relationship between arrival rules, linear partial balance equations, and product form stationary distributions. This model is motivated by production systems operating under a kanban protocol. In contrast with the conventional departure first networks, where a transition is initiated by service completion of items at the originating nodes that are subsequently routed to the destination nodes (push system), in an arrival first network a transition is initiated by the destination nodes of the items and subsequently those items are processed at and removed from the originating nodes (pull system). These are similar to the push and pull systems in manufacturing systems.

Our characterisation provides necessary and sufficient conditions for the network to possess linear traffic equations, and sufficient conditions for the network to have a product form stationary distribution. We apply our results to networks operating under a kanban mechanism and characterise the rate at which items are pulled as well as the routing and blocking protocols that give rise to a product form stationary distribution.     

An agent-based support system for railway station dispatching

For those railway stations without being automated, railway traffic dispatching still depends on dispatchers, especially under disturbed circumstances. In this study, an agent-based support system, named D-Agent, is developed to assist human dispatchers to make decisions in station operation. To this end, the common knowledge and possible difficulties concerning a station dispatcher in his/her routine work are firstly studied, and the D-Agent is proposed with the purpose of working out practicable solutions to these challenging tasks as a dispatcher does. Then the general model of the D-Agent is established, containing five basic modules: local database, knowledge base, skill base, reasoning mechanism and communication interfaces. The internal skills of the D-Agent are designed to execute various tasks in different scenarios. Besides, a skill extension of the D-Agent with mathematical formulations is particularly discussed in this paper, to find feasible and optimal traffic control solutions in disturbance situations such as train delays and route conflicts. The D-Agent is designed to learn from its own experimental history in applying different skills, and evaluate the skills by preference weights of alternative solutions in a particular task. This procedure allows the agent to have potential for continuous improvement. To verify the applicability of the proposed support system, a D-Agent for a terminal station of subway is simulated. The numerical example of train delays and route conflicts shows that the D-Agent can generally perform as a station dispatcher in fulfilling the specific tasks, estimate the traffic state in different operation strategies and support the decision-making of favored solutions. Significantly, it indicates that the mathematical methods can also been employed by an intelligent agent.     

基于竞争终端个数区间的IEEE 802.11性能优化

IEEE 802.11的MAC协议采用基于CSMA/CA的DCF机制,研究发现,上述协议的性能随无线局域网中竞争终端个数的增加而迅速恶化.当竞争终端个数在一定范围内变化时,使用相同的优化协议参数,系统的性能都能接近最优.因此,设计了一个基于状态检测与竞争终端个数区间的自适应性能优化机制,DOOR(dynamicootimization on range).根据相关性能模型的分析,先将竞争终端的个数分为若干区间,并分别计算出各区间的优化协议参数.当系统检测到竞争终端个数发生变化时,根据其所处的区间,对相关的参数进行动态调整,从而有效地改善了协议的整体性能.同时还给出了相关理论模型和计算的详细说明,并尝试给出了划分区间的基本原则与方法.最后,实验仿真结果验证了新的方法能够根据竞争终端个数的变化对系统性能进行整体优化,在吞吐量和延迟等方面明显优于标准的IEEE 802.11协议.     

Robust manufacturing system design using multi objective genetic algorithms,Petri nets and Bayesian uncertainty representation

Decisions involving robust manufacturing system configuration design are often costly and involve long term allocation of resources. These decisions typically remain fixed for future planning horizons and failure to design a robust manufacturing system configuration can lead to high production and inventory costs, and lost sales costs. The designers need to find optimal design configurations by evaluating multiple decision variables (such as makespan and WIP) and considering different forms of manufacturing uncertainties (such as uncertainties in processing times and product demand). This paper presents a novel approach using multi objective genetic algorithms (GA), Petri nets and Bayesian model averaging (BMA) for robust design of manufacturing systems. The proposed approach is demonstrated on a manufacturing system configuration design problem to find optimal number of machines in different manufacturing cells for a manufacturing system producing multiple products. The objective function aims at minimizing makespan, mean WIP and number of machines, while considering uncertainties in processing times, equipment failure and repairs, and product demand. The integrated multi objective GA and Petri net based modeling framework coupled with Bayesian methods of uncertainty representation provides a single tool to design, analyze and simulate candidate models while considering distribution model and parameter uncertainties.     

Near-optimal broadcast in all-port wormhole-routed hypercubes usingerror-correcting codes

A new broadcasting method is presented for hypercubes with wormhole routing mechanism. The communication model assumed allows an n-dimensional hypercube to have at most n concurrent 110 communications along its ports. It further assumes a distance insensitivity of (n+1) with no intermediate reception capability for the nodes along the communication path. The approach is based on determination of the set of nodes (called stations) in the hypercube such that for any node in the network there is a station at distance of at most 1. Once stations are identified, parallel disjoint paths are formed from the source to all stations. The broadcasting is accomplished first by sending the message to all stations which will in turn inform the rest of the nodes of the message. To establish node-disjoint paths between the source node and all stations, we introduce a new routing strategy. We prove that multicasting can be done in one routing step as long as the number of destination nodes are at most n in an n-dimensional hypercube. The number of broadcasting steps using our routing is equal to or smaller than that obtained in an earlier work; this number is optimal for all hypercube dimensions n⩽12, except for n=10     

WIPA: neural network and case base reasoning models for allocating work in progress

Assembly Line Balance (ALB) problem is a typical combinatorial optimization problem where pieces of work are transported between the work stations. In the ALB problem, the ultimate goal is to seek the optimal makespan. It is a very difficult problem to solve particularly in Sewn Product Industry (SPI) which is a labor-intensive manufacturing industry. In order to achieve the optimal makespan, it is necessary to take into account factors such as the efficiency of each machinist, the allocation of suitable Work In Progress (WIP) into each assembly line, the calculation of each product production time in terms of Sewing Minute Value (SMV) and the assignment of each machinist into different work stations according to his/her capability. However, the current methodologies are dependent on human experts relying on statistical data. These data, however, are problematic in that they are historical data and as such are unlikely to be suitable for all circumstances especially as in a highly competitive industry such as the SPI practices, standards and tasks are constantly changing and adapting. In this paper, two models have been proposed to solve the WIP allocation problem and the SMV calculation problem. The preliminary results are encouraging. The first model is able to extract a large number of the rules and has attained a prediction accuracy of 93%. The second model can increase 11% in accuracy in predicting the SMV compared to the current widely used General Sewing Data (GSD) method.     

考虑多维修台异步多重休假的温贮备冗余系统可靠性模型

考虑多维修台保障多个系统时维修力量存在调度与分配的情况,引入多维修台异步多重休假策略;以温贮备冗余系统为研究对象,针对以往研究利用指数分布等典型分布导致模型约束条件过于严格的问题,采用连续phase-type(PH)分布描述系统中工作部件寿命、温贮备部件寿命以及维修台休假时间和维修时间,建立通用性更好的系统可靠性解析模型,给出系统可靠度、系统稳态可用度等冗余系统可靠性指标和稳态忙期维修台数量等维修台稳态指标;利用算例验证模型适用性,演示了维修台数量、系统温贮备部件数量变化以及修理工休假速率、维修速率变化对系统各可靠性指标和维修台稳态指标的影响.算例计算结果表明,所提出的可靠性模型能够有效复现多维修台调度对冗余系统可靠性的影响,从而为维修台数量的合理安排及系统部件数量的优化配置提供理论基础和实践参考.     

Design of bi-criteria kanban system using simulated annealing technique

In the kanban system, the main decision parameters are the number of kanbans and lot size. In this paper, an attempt has been made to set the number of kanbans at each station and the lot size required to achieve the best performance using simulated annealing technique. A simulation model with a single-card system has been designed and used for analysis. A bi-criterion objective function comprising of mean throughput rate and aggregate average kanban queue has been used for evaluation. Different perturbation schemes have been experimented and compared.     

A decision support system for design of mass production machining lines composed of stations with rotary or mobile table

This paper describes a decision support system (DSS) developed in order to offer to machining line designers a cognitive aid for early design stages. The aim of DSS is to assist the decision makers in finding the configuration of a new line that will meet quality and productivity requirements and minimize the investment costs. The current version of DSS is oriented to design of mass production machining lines composed of machines with rotary or mobile tables. This decision support system is based on mathematical models and methods which were devised to provide the designers with the optimal parameters of new line configuration including the required number of working stations of different types, the number of working positions at each station and spindle heads at each working position. The system is implemented under Autodesk Inventor and includes the modules for part modeling, process planning and machining system configuration. Its modular character and open architecture make upgrading with new mathematical tools suitable for other machining systems easy and fast. Moreover, it can be employed either as a separate software or integrated in a Product Life-cycle Management (PLM) tool.     

Assembly line balancing with positional constraints,task assignment restrictions and station paralleling: A case in an electronics company

In this paper, we present a real-life Assembly Line Balancing Problem for an electronics manufacturing company. The main characteristics of the problem are as follows: (i) a set of operations are related to the front part of the workpiece and others are related to the back part of the workpiece, which in turn makes all tasks dependent on the position of the workpiece, (ii) some of the tasks must be executed on the same station and no other tasks should be assigned to this station due to technological restrictions, (iii) parallel stations are allowed to increase the line efficiency at the required production rate and to overcome the problem of assigning tasks with operation times that exceed the cycle time. Initially, the problem is formulated as a 0–1 integer programming model and solved using CPLEX solver. Then, the effect of alternative work schedules such as multiple shifts and overtime on the expected labor cost of the line is analyzed. Considering alternative work schedules while balancing the line for corresponding cycle times allows us to select an efficient assembly line for the company, resulting in a lower labor cost and a more balanced line with respect to the operation times and the activity of the workers at each station. Lastly, a computational study is conducted to evaluate the performance of the proposed model. It is found that the model is capable of producing high quality solutions in reasonable solution times.     

Single and Multi-objective Evolutionary Algorithms for the Coordination of Serial Manufacturing Operations

This paper focuses on a typical problem arising in serial production, where two consecutive departments must sequence their internal work, each taking into account the requirements of the other one. Even if the considered problem is inherently multi-objective, to date the only heuristic approaches dealing with this problem use single-objective formulations, and also require specific assumptions on the objective function, leaving the most general case of the problem open for innovative approaches. In this paper, we develop and compare three evolutionary algorithms for dealing with such a type of combinatorial problems. Two algorithms are designed to perform directed search by aggregating the objectives of each department in a single fitness, while a third one is designed to search for the Pareto front of non-dominated solutions. We apply the three algorithms to considerably complex case studies derived from industrial production of furniture. Firstly, we validate the effectiveness of the proposed genetic algorithms considering a simple case study for which information about the optimal solution is available. Then, we focus on more complex case studies, for which no a priori indication on the optimal solutions is available, and perform an extensive comparison of the various approaches. All the considered algorithms are able to find satisfactory solutions on large production sequences with nearly 300 jobs in acceptable computation times, but they also exhibit some complementary characteristics that suggest hybrid combinations of the various methods.