International Journal of Industrial Engineering: Theory, Applications and Practice
Not a member yet
943 research outputs found
Sort by
CHALLENGES OF DISCRETE EVENT SIMULATION IN THE EARLY STAGES OF PRODUCTION SYSTEM DESIGN
No full textThis study analyzes the challenges of applying discrete event simulation in the early stages of production system design. Highlighting the implications of new production processes and technologies leading to improved competitiveness, this study provides novel contributions to the understanding of discrete event simulation based on three case studies of the transformation of legacy production systems in the heavy vehicle industry. The findings of this study show that equivocal or ambiguous understanding about new production processes or technologies, and uncertainty about necessary data input and the interrelation of subsystems in production, are critical in addressing discrete event simulation-related challenges. These findings highlight the need for an established process to manage assumptions and simplifications during the design, development, and deployment of discrete event simulation models as a countermeasure against uncertainties, improving manufacturing system design and practice
VULNERABILITY ANALYSIS FOR EVACUATION TRANSPORTATION NETWORKS
No full textAs occurrences of natural disasters are increasing, interests of reacting for the certain circumstances such as evacuation planning have been rising. When an evacuation network is built, strategic analyses to support operation of the network are needed. Vulnerability analysis that seeks for weak points of an evacuation transportation network is one of those. This paper aims to analyze vulnerability of evacuation transportation network constructed with the Cell Transmission Model (CTM). Previous researches have considered all the possible scenarios to investigate weak points without any mathematical approaches. Accordingly, we propose a bi-level optimization model to identify fragile spots of the network. Due to a complexity of the problem, genetic algorithm combined with a linear programming is developed to find good solutions in reasonable computational time. Solutions from the proposed algorithm are compared with optimal solutions
PREDICTION MODEL BASED MULTI-PROFILE MONITORING FOR MANUFACTURING PROCESS MANAGEMENT
No full textIn an advanced manufacturing environment, the analysis of profile data collected from the process equipment is a critical issue in improving process efficiency. In particular, multi-profile monitoring is essential for process control because an advanced manufacturing process consists of numerous pieces of equipment and their related sensors. The main goal of this study is to build a monitoring chart using a Profile Integrated Measure (PIM) from multi-profile data in order to observe an overall condition of various points in the process. To deploy the proposed algorithm, multi-profile data needed to be preprocessed and applied to the prediction model. The PIM is calculated from the prediction model and reflects the relationships between the multi-profile data property, which has normal/abnormal states. The proposed algorithm constructs a model using the PIM of a normal state and identifies the performance of the model. Experiments with the simulation datasets modified from the manufacturing process validate the effectiveness and applicability of the proposed algorithm
A Study of Software Reliability Growth with Imperfect Debugging for Time-Dependent Potential Errors
No full textOver the last few decades, various software reliability growth models (SRGM) have been proposed, and in recent years a gradual but marked shift has been focused on the balance between acceptable software reliability and affordable software testing cost. Chiu et al.(2008) proposed an SRGM from the perspective of learning effects which is more flexible in fitting various software error data, although it has a restrictive assumption of constant number of potential errors. In this paper, we consider a software reliability growth model in which the number of potential errors varies over the debugging period, since wrongly fixing an error may cause more errors; while correctly debugging one error may resolve others. However, such variation would gradually converge as the testing staff becomes more familiar with the software system. In order to describe this phenomenon, a sine function is introduced in this study to describe the time-dependent behavior of the number of potential errors with imperfect debugging, and the expected testing cost is thus evaluated to assist in the determination of an optimal software release policy. A numerical example is illustrated to verify the effectiveness of the proposed approach
A PRODUCTION LOT SIZING MODEL CONSIDERING QUALITY CONTROL, REWORK, AND SHORTAGES
No full textThis paper investigates a production lot sizing problem where defective items produced during a production run can be reworked after inspections conducted during and at the end of production. The rework process is considered to produce a fixed proportion of scrapped items. Shortages are allowed and completely backordered. In process industries such as semiconductor manufacturing, understanding of the influence of rework on inventory is needed for effective and controlled planning. The aim of this study is to determine the optimal production quantity and maximum shortage level to maximize the expected total profit per unit of time. A mathematical model is presented along with a proof of the concavity of the expected total profit function. A numerical example is presented to illustrate the effectiveness of the proposed model, as well as a sensitivity analysis of optimal solutions with respect to major parameters
PREDICTION OF TIMES-TO-FAILURE OF SEMICONDUCTOR CHIPS USING VMIN DATA
No full textAccelerated Life Testing (ALT) aims at predicting times-to-failure under normal operating condition. The prediction requires times-to-failure data under ALT operation conditions, however, it is difficult to obtain the times-to-failure data of semiconductor chips when only few failures occur. In this regard, we attempt to predict times-to-failure of semiconductor chips by using Vmin data. Since Vmin are measured for all of tested chips regardless of failure, we can predict times-to-failure for all of the chips. The proposed method is more informative and robust than the traditional life data approach in that all of the tested semiconductor chips participate in the life prediction proces
A hybrid estimation of distribution algorithm for a container pre-marshalling process
No full textIn a yard storage area, pre-marshaling means relocating the export containers into a proper arrangement in order to increase the efficiency of the load process. The container pre-marshaling process aims to identify the best sequence of movements of container movements on an initial layout of the yard storage area, so that it reaches a desired final layout and satisfies operational constraints, while minimizing the total number of movements required to achieve the best sequence. Contrary to current research, a new evolutionary algorithm is proposed and developed to solve the pre-marshaling problem and simulate the solution. Our approach combines the key advantages of both evolutionary algorithms and the Mallows model. The Mallows distribution is used to model the pre-marshaling scenario, while an evolutionary algorithm is used to guide the overall search process to identify the best performing sequences. The approach makes use of the Mallows model to describe the distribution of the solution space. The proposed algorithm is able to identify the next most probable movement in the yard storage area. General and standard benchmarking and real-world cases served as input and test parameters in order to show the performance of the proposed algorithm
PRODUCT DESIGN TIME FORECAST USING RELATIVE ENTROPY KERNEL REGRESSION
No full textDesign time forecast is characterized by problems of small samples and heteroscedastic noise. For better control of product development process, the forecast method is expected to yield not only highly precise forecast values but valid forecast variances. For this end, relative entropy kernel regression (REKR) is proposed based on the combination of kernel-based regression and Gaussian margin machines (GMM). REKR maintains a Gaussian distribution over weight vectors for the kernel-based regression to minimize the combination of the relative entropy and the negative log probability densities of the targets. To simplify the formulation of REKR, the covariance matrix of Gaussian distribution is set as a diagonal matrix. The simplified optimization problem of REKR is solved based on particle swarm optimization (PSO). Having inherited the benefits of GMM, REKR can simultaneously offer the desirable forecast value and forecast variance. Experiments on the time forecast of plastic injection mold design as well as on both synthetic and real datasets do verify the feasibility and validity of REKR
MULTI-PERIOD SUPPLY CHAIN COORDINATION USING TRADE PROMOTION: COMPLEMENTARY SLACKNESS APPROACH
No full textIn this paper, a two-level supply chain network with a single manufacturer supplying a single product to a single retailer is studied. This research uses a trade promotion strategy to coordinate the supply chain by finding the optimal pre-announced multi-period wholesale prices that can induce the retailer’s decentralized decisions to be the same as the retailer’s centralized decisions with the minimum total cost for the supply chain. The manufacturer makes production, inventory, and wholesale price decisions. The retailer makes ordering and inventory decisions. A procedure is proposed to determine optimal wholesale prices to pre-announce in each period to the retailer, coordinating the supply chain using complementary slackness conditions. The results show the coordination benefits for a supply chain when the setup or reorder cost is high but the average demand is low. Finally, the performance of the proposed method is compared with the performance of the “Every Day Low Price” wholesale price policy
A COMPLETE DESIGN METHODOLOGY FOR LEAN IN-PLANT LOGISTICS TO ASSEMBLY LINE USING AD PRINCIPLES
No full textImplementation of lean philosophy has gained importance since it has achieved an outstanding success especially in production systems. Principles of lean production are different from traditional systems, where parts feeding and material handling activities are performed differently especially in small lot production. Integration and coordination of these activities increase the efficiency of lean systems. In this paper, a complete design for in-plant logistics activities is performed via Axiomatic Design (AD) principles. A four-step design methodology from parts line side presentation through in-plant vehicle routing is proposed. Comprehensive literature for each design step is given, and a continuous improvement scheme for the proposed design is developed. The relations between lean principles, design parameters and lean tools are evaluated and the justification of the proposed design is shown. The findings of the study also point out voids in the literature and shed light on future studies.