Kanban System Research Articles

Evaluating artificial intelligence heuristics for a flexible Kanban system: simultaneous Kanban controlling and scheduling

This paper evaluates several artificial intelligence heuristics for a simultaneous Kanban controlling and scheduling on a flexible Kanban system. The objective of the problem is to minimise a total production cost that includes due date penalty, inventory, and machining costs. We show that the simultaneous Kanban controlling and scheduling is critical in minimising the total production cost (approximately 30% cost reduction over scheduling without a Kanban controlling). To identify the most effective search method for the simultaneous Kanban controlling and scheduling, we evaluated widely known artificial intelligence heuristics: genetic algorithm, simulated annealing, tabu search, and neighbourhood search. Computational results show that the tabu search performs the best in terms of solution quality. The tabu search also requires a much less computational time than the genetic algorithm and the simulated annealing. To further improve the solution quality and computational time for a simultaneous Kanban controlling and scheduling on a flexible Kanban system, we developed a two-stage tabu search. At the beginning of the tabu search process, an initial solution is constructed by utilising the customer due date information given by a decision support system. The two-stage tabu search performs better than the tabu search with a randomly generated initial solution in both the solution quality and computational time across all problem sizes. The difference in the solution quality is more pronounced at the early stages of the search.

Design of multi-stage adaptive kanban system

The traditional kanban system with a fixed number of cards does not work satisfactorily in an unstable environment. In the adaptive kanban-type pull control mechanism, the number of kanbans are 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 local search and meta-heuristic methods to design the adaptive kanban system for single stage. In a multi-stage system the cards are circulated within the stage and their presence at designated positions will signal to the neighboring stages about the inventory. In this work, a model of the multi-stage system to traditional and adaptive kanban system is developed. A GA-based search is employed to set the parameters of the system. The results are compared with a traditional kanban system and found signs of improvement.

Production à flux tirés dans une chaîne logistique

The choice of a production management between a pull-flow approach and a push-pull approach is less simple than it seems when we take into account the perimeter of the logistics chains. In fact, it is generally necessary to combine these two management methods in order to meet increasingly short customer deadlines as well as possible without increasing costs. These couplings are not independent of the characteristics of the logistics chain, of the products to be manufactured and of the regularity of demand, and calls for different forms of management. We will examine here one of the most well-known approaches, the kanban system, which has been used successfully under certain conditions, in particular that of the stability of demand. We will analyze the use of the kanban system in environments violating the assumptions conventionally made in this field, which will make it possible to highlight the conditions of robustness of this type of piloting. The questioning of the cruising regime in the kanban system somewhat complicates the analysis of the modalities of flow management on the logistics chain and pushes to imagine new management rules to allow the synchronization of productions in the logistics chain.

Analysis of decentralized multi‐product pull systems with lost sales

AbstractWe study a pull‐type, flexible, multi‐product, and multi‐stage production/inventory system with decentralized two‐card kanban control policies. Each stage involves a processor and two buffers with finite target levels. Production stages, arranged in series, can process several product types one at a time. Transportation of semi‐finished parts from one stage to another is performed in fixed lot sizes. The exact analysis is mathematically intractable even for smaller systems. We present a robust approximation algorithm to model two‐card kanban systems with batch transfers under arbitrary complexity. The algorithm uses phase‐type modeling to find effective processing times and busy period analysis to identify delays among product types in resource contention. Our algorithm reduces the effort required for estimating performance measures by a considerable margin and resolves the state–space explosion problem of analytical approaches.Using this analytical tool, we present new findings for a better understanding of some tactical and operational issues. We show that flow of material in small procurement sizes smoothes flow of information within the system, but also necessitates more frequent shipments between stages, raising the risk of late delivery. Balancing the risk of information delays vis‐à‐vis shipment delays is critical for the success of two‐card kanban systems. Although product variety causes time wasted in setup operations, it also facilitates relatively short production cycles enabling processors to switch from one product type to another more rapidly. The latter point is crucial especially in high‐demand environments. Increasing production line size prevents quick response to customer demand, but it may improve system performance if the vendor lead‐time is long or subject to high variation. Finally, variability in transportation and processing times causes the most damage if it arises at stages closer to the customer. © 2007 Wiley Periodicals, Inc. Naval Research Logistics, 2007

Comparison of supply chain replenishment strategies in a non-capacitated distribution system

This research considers inventory replenishment in a stochastic, non-capacitated, multi-echelon distribution system. Simulation is used to model Distribution Requirements Planning (DRP), reorder point and Kanban replenishment strategies. Additional experimental factors include demand seasonality, lot sizes, demand uncertainty and transit time uncertainty. Trade-off curves between inventory and delivery performance are generated. Statistical techniques, including ANOVA and regression, are then used to determine the relative dominance of the replenishment strategies by comparing the areas under the trade-off curves. Results show the methodology developed for analysis of the results works well in identifying both main and interaction effects. DRP performance is found to be the most robust, followed by that of reorder point and Kanban systems, respectively.

New Erase Head for RFID Thermal Rewritable Media

Although thermal rewritable technology has been known in the printing industry for about a decade, it has been used mainly for card applications only in Japan and Europe. Usages in other applications have been less than anticipated for various reasons.Radio Frequency Identification (RFID) technology has been around much longer than rewritable technology, but its usages have been becoming popular in the last several years.A new application is emerging which utilizes both of those technologies in industrial manufacturing operation known as “just-in-time” (JIT) Kanban System. Using RFID for Kanban media makes the operation more efficient as the information on the card can be retrieved beyond bar-code scanning range and electronic data on the media can be updated. One issue with electronic data on the media (card or label) update is that the media needs to be reprinted if the visible information (such as alpha-numeric and bar-code) needs to be updated while electronic data can be changed easily and almost endlessly. Regular media with memory IC chip is rather expensive and single use (print-and-discard) process is uneconomical. This is where the rewritable technology comes in, since the media can be reused (rewritten) several hundred times – the cost becomes almost a non-issue.HIT Devices, Ltd., developed an erase head specifically designed to meet the requirements of this application.

Performance analysis of single stage kanban controlled production systems using parametric decomposition

We present a new approach that permits efficient performance analysis of kanban systems with general demand processes, material arrival processes, and service times. The approach is based on parametric characterization of the traffic processes (arrival and departure) in the network and uses two-moment approximations to estimate performance measures at individual stations. We derive traffic flow constraints that are particular to closed queuing networks with synchronization stations and use these to establish relationships between the parameters characterizing arrival and departure processes at the stations in the network. The resultant set of non-linear equations is solved to estimate network performance measures. Numerical studies show that the approach is not only fast but also reasonably accurate when compared to simulation. These studies also provide insights with respect to the impact of different types of variability on the performance of a kanban system. This work also provides a fundamental building block that can be used in the analysis of multi-stage kanban systems.

THE KANBAN SYSTEM FOR THE MANUFACTURE OF PAINT

This paper presents a case study for the project and implementation of a kanban system in a paint factory. The paint factory uses a continuous production system, which brings two main problems for the kanban system. The first is that of the number of kanbans in the system (that is, the determination of the amount of each batch to be manufactured). The formula which has been consecrated by Toyota requires that the manufacture time of each batch is known, In addition, in the case in question, the entire batch is finished in a single period. The second problem is that of altering the size of the batch during the manufacture. In discrete production systems (car parts, for example), as the intention is to avoid the accumulation of stock, it is usual to interrupt the production of an item as soon as all the empty boxes are filled again. In this case, if, at the beginning of the manufacture of an item there are kanbans still in stock, this decision is difficult if one wishes to avoid the lack of excess stock in the system. This study presents a proposal to deal with these problems.

Literature review of JIT-KANBAN system

In this paper, JIT (Just-In-Time)-KANBAN literature survey was carried out and presented. The introductory section deals with the philosophy of JIT, and the concept involved in the push and pull system. The blocking mechanisms in the kanban system are also discussed elaborately. Besides these sections, the importance of measure of performance (MOP) and the application of the same with respect to JIT-KANBAN are presented. The recent trends in the JIT-KANBAN are discussed under the heading “Special cases”. In this review, 100 state-of-art research papers have been surveyed. The directions for the future works are also presented.

Research and Issues on Kanban Systems

Research and Issues on Kanban Systems

EVOLUTION OF KANBAN SYSTEMS THANKS TO A (MAX,+)-ALGEBRA ANALYSIS

This paper deals with a possible evolution of a Kanban system due to a (max,+)-algebra analysis. We show that for a given Kanban system, it is always possible to change the original Kanban policy by a (max,+)-linear policy which keeps the same quality of service but reduces the work in process. This new control policy contains a (max,+)-linear dynamic behavior for the recycling of kanban cards.

The application of analytic hierarchy process in classification of material planning and control systems

Many organisations keep thousands of items in their warehouses, which may worth more than 50% of all companies' expenses (Neely and Byrne, 1992). Controlling of these items by one Material Planning and Control System (MPCS) (push and pull systems) may not be rational or economic. Every methodology has different cost and service level in different environments. Therefore, the tool required to categorise the material based upon the suitability of MPCS. This paper presents a multicriteria approach within Analytic Hierarchy Process (AHP), which is capable of classifying the materials and components for different MPCS, that is, Kanban, MRP, hybrid and Re-Order Point (ROP) Systems.

In Lean Production: Semantics Matters

Manufacturing, Planning, and Control (MPC) systems are commonly categorized as either push or pull production systems. The most common example of a push system ref erred to is a MRP system, while the example of a pull system most commonly ref erred to is JIT, or more specifically a Kanban system (Sawaya et al, 1992). We have found that the use of this terminology regarding different forms of MPC systems can be confusing to the novice as well as the more experienced management practitioner or researcher. In this paper, we make an argument for the importance of using more precise MPC terms and concepts to foster a more universally consistent understanding of MPC systems. In response to inconsistencies of definitions, explanations, and examples, we provide a framework for classifying MPC systems; a brief discussion of the types of performance measures to monitor or control for effective performance and continuous improvement in either a push or pull system.

So many factors, so little time…Simulation experiments in the frequency domain

We illustrate how an efficient methodology called frequency domain experimentation can be used to gain better insight into the behavior of production systems. With the full factorial designs commonly used for simulation experiments, the number of runs grows exponentially with the number of factors involved, while the run length remains constant. In frequency domain experiments, the number of runs is independent of the number of factors, while the required run lengths increase relatively slowly. We describe the method, illustrate its effectiveness at identifying important main effects, two-way interactions, and quadratic terms for a known model, demonstrate the approach by evaluating a kanban system involving 34 factors, and provide links to software. We also present computational requirements for running simulation experiments that combine a batch means approach with efficient run-oriented designs for a variety of systems. The results indicate that frequency domain experiments perform very favorably for systems, such as queueing networks, where the simulation's output stream exhibits high positive autocorrelation.

Analysis of assembly systems controlled with kanbans

This paper deals with the problem of evaluating the performance of assembly systems managed with kanbans. In particular we analyze kanban systems functioning with different control policies depending on how kanbans are released in assembly stages. Simultaneous and independent releases of kanbans are considered in the analysis as alternative control policies to be used in assembly systems. Queuing network techniques are used to calculate the major system performance measures such as throughput of the system, percentage of satisfied orders, average level of finished products and average time delay of unsatisfied orders. Simultaneous and independent kanban assembly systems are compared in some numerical cases. The differences, in terms of performance, between the two alternative kanban control policies are evaluated so that the proper control policy can be selected in the design phase of the system. Simulation results are also reported to validate the approximation of the analytical method used to calculate the performance of assembly kanban systems.

Kanban implementation at a tyre manufacturing plant: a case study

This paper deals with the implementation of the Kanban system in a continuous product line of a tyre manufacturing plant. The paper also describes a practical approach to the design of Kanban card, board, container sizing, the number of Kanban, operating rules and day-to-day scheduling of machines using Kanban, and the necessary changes in a wage system to support its acceptability by workers on the shop floor.

Synergy between evolutionary optimization and induction graphs learning for simulated manufacturing systems

The performance of manufacturing systems is strongly dependent on a set of numerical and non-numerical decision variables, such as transport lot sizes and priority rules, for example. Configuring such systems requires determining a value for each decision variable to optimize a performance criterion. Recent works have shown that simulation optimization of manufacturing systems can be efficiently addressed using evolutionary algorithms. Nevertheless, these algorithms do not provide any understanding on the system's behaviour, which could be crucial regarding the investment amounts. Existing methods based on learning strategies try to provide knowledge on the system, but relevant information could be scattered. To optimize the system, characterize good solutions and highlight critical decision variables, an ‘intelligent optimization’ method is proposed based on the synergy between evolutionary optimization and induction graphs learning. These approaches work together to find good solutions with associated knowledge. The method is illustrated by the configuration of an assembly kanban system constituted of five machines that process three types of product.

Operation Method for Multi-Types and Multi-Stages Kanban System Using Discrete Event Simulation

Many studies have been done to optimize the Kanban system as production control system. However, most of them can be effective on only limited conditions. Therefore, these studies have difficulties to be applied in practice. The purpose of this study is to propose the practical method to operate Multi-types and Multi-stages Kanban system using discrete event simulation. The method sets the number of Kanbans much larger than the estimated number on the simulation model. From the simulation result, the surplus inventory is found. By subtracting the surplus inventory from the maximum inventory, the necessary number of Kanbans and the average of inventory are determined. Also the changeover time is calculated from the simulation result. The performance of Kanban system can be measured through the average of inventory and the changeover time. The best plan to operate Kanban system is selected among many alternatives based on that performance. To confirm the effectiveness of the method, an actual test case is tried out.

The Kanban System's Permeation within Toyota

The Kanban System's Permeation within Toyota

Reactive JIT ordering system for changes in the mean and variance of demand

This paper proposes a reactive just-in-time (JIT) ordering system for multi-stage production systems with unstable changes in demand, not only in the mean but also in the variance. For a JIT ordering system, that is, the Kanban system, a reactive controller of buffer size is proposed. In the proposed system, unstable changes in the mean and variance of demand can be detected by exponentially weighted moving average charts for mean and variance, and the buffer size at each inventory point is controlled as a reaction to the detected unstable change. The performance of the proposed system is analyzed by simulation experiments.