- +886-3-5742150
Research Interests:
Research Interests:
The re-entrant flow shop scheduling problem considering time windows constraint is one of the most important problems in hard-disc drive (HDD) manufacturing systems. In order to maximise the system throughput, the problem of minimising... more
The re-entrant flow shop scheduling problem considering time windows constraint is one of the most important problems in hard-disc drive (HDD) manufacturing systems. In order to maximise the system throughput, the problem of minimising the makespan with zero loss is considered. In this paper, evolutionary techniques are proposed to solve the complex re-entrant scheduling problem with time windows constraint in manufacturing HDD devices with lot size. This problem can be formulated as a deterministic Fm | fmls, rcrc, temp | Cmax problem. A hybrid genetic algorithm was used for constructing chromosomes by checking and repairing time window constraints, and improving chromosomes by a left-shift heuristic as a local search algorithm. An adaptive hybrid genetic algorithm was eventually developed to solve this problem by using fuzzy logic control in order to enhance the search ability of the genetic algorithm. Finally, numerical experiments were carried out to demonstrate the efficiency of the developed approaches.
Research Interests:
ABSTRACT Effective supply chain management (SCM) comprises activities involving the demand and supply of resources and services. Negotiation is an essential approach to solve conflicting transaction and scheduling problems among supply... more
ABSTRACT Effective supply chain management (SCM) comprises activities involving the demand and supply of resources and services. Negotiation is an essential approach to solve conflicting transaction and scheduling problems among supply chain members. The multi-agent ...
Research Interests:
Research Interests:
Research Interests:
The Atacama Large Millimeter/submillimeter Array (ALMA) Band 1 receiver covers the frequency range of 35-50 GHz. An extension of up to 52 GHz is on a best-effort basis. A total of 73 units have to be built in two phases: 8 preproduction... more
The Atacama Large Millimeter/submillimeter Array (ALMA) Band 1 receiver covers the frequency range of 35-50 GHz. An extension of up to 52 GHz is on a best-effort basis. A total of 73 units have to be built in two phases: 8 preproduction and then 65 production units. This paper reports on the assembly, testing, and performance of the preproduction Band 1 receiver. The infrastructure, integration, and evaluation of the fully-assembled Band 1 receiver system will be covered. Finally, a discussion of the technical and managerial challenges encountered for this large number of receivers will be presented.
Research Interests:
This study aims to develop the UNISON Framework as a systematic approach for decision analysis that can assist the decision maker in the comprehensive processes for defining and structuring the right decision problem,clarifying the... more
This study aims to develop the UNISON Framework as a systematic approach for decision analysis that can assist the decision maker in the comprehensive processes for defining and structuring the right decision problem,clarifying the decision elements,collecting the data,extracting useful information,evaluating the alternatives to select the optimal decision for effective decision support.The proposed UNISON framework can enhance the decision quality of complex decision problems with justification.In this paper,the six phases of UNISON framework are explained in details and specific case studies are used to illustrate its validity.
Research Interests:
Recently, unmanned aerial vehicles (UAVs) enhance connectivity and accessibility for civilian and military applications. A group of UAVs with on-board cameras usually monitors or collects information about designated areas. The UAVs can... more
Recently, unmanned aerial vehicles (UAVs) enhance connectivity and accessibility for civilian and military applications. A group of UAVs with on-board cameras usually monitors or collects information about designated areas. The UAVs can build a distributed network to share/exchange and to process collected sensing data before sending to a data processing center. A huge data transmission among them may cause latency and high-energy consumption. This paper deploys artificial intelligent (AI) techniques to process the video data streaming among the UAVs. Thus, each distributed UAV only needs to send a certain required information to each other. Each UAV processes data utilizing AI and only sends the data that matters to the others. The UAVs, formed as a connected network, communicate within a short communication range and share their own data to each other. Convolution neural network (CNN) technique extracts feature from images automatically that the UAVs only send the moving objects i...
Research Interests:
Research Interests:
Silicon wafers are critical raw materials for semiconductor fabrication. Wafer characteristics and specifications will affect the yield of integrated circuits fabricated on the wafer. As critical dimensions for semiconductor manufacturing... more
Silicon wafers are critical raw materials for semiconductor fabrication. Wafer characteristics and specifications will affect the yield of integrated circuits fabricated on the wafer. As critical dimensions for semiconductor manufacturing are shrinking, defining wafer characteristics and specifications for the corresponding semiconductor devices is crucial for yield enhancement and smart manufacturing. However, to the best of our knowledge, none of existing studies have investigated the relationship between wafer characteristics and process variables. Focusing on the needs in real settings, this study aims to develop an advanced quality control (AQC) solution for raw wafers for yield enhancement in advance, in which the UNISON framework is employed to derive the optimal raw material specifications for different products. For validation, an empirical study was conducted in a leading semiconductor fab to derive rules to effectively improve the quality indices including defect count, overlay errors, anti-contamination capability, and yield. The results have shown practical viability of the proposed AQC approach as associated effort to enhance existing approaches including advanced process/equipment control (APC/AEC). Indeed, the developed solution is implemented in real settings as partial effort for enabling Industry 3.5.
Research Interests:
Research Interests:
Research Interests:
Research and development (R&D) projects are crucial for semiconductor companies to maintain growth, profitability, and competitiveness. Integrated circuit (IC) design is capital intensive and continuously migrates to new technologies... more
Research and development (R&D) projects are crucial for semiconductor companies to maintain growth, profitability, and competitiveness. Integrated circuit (IC) design is capital intensive and continuously migrates to new technologies to meet various market demands. Moreover, the scheduling of selected R&D projects that enables technology roadmap involving complicated interrelationships, while competing for similar resources. Focusing on realistic needs, this paper aims to propose an integrated approach for selecting IC design projects for R&D portfolios and scheduling the selected projects simultaneously. In particular, a hybrid autotuning multiobjective genetic algorithm was developed to solve large sized problem instances. An empirical study was conducted at a leading IC design service company in Taiwan to test the validity of the proposed approach. The proposed algorithm was compared with conventional approaches for both convergence and diversity. The results have shown the practical viability of this approach in efficiently and effectively generating near-optimal portfolio alternatives for portfolio selection. The approach also enables the scheduling of the selected projects to achieve R&D portfolio objectives. The developed solution was fully implemented and adopted by the company.
Research Interests:
Research Interests:
A data mining and data-driven framework is constructed to extract user preferences effectively.An empirical study was conducted to derive useful rules for product form of wearable devices.Specific rules are employed to support product... more
A data mining and data-driven framework is constructed to extract user preferences effectively.An empirical study was conducted to derive useful rules for product form of wearable devices.Specific rules are employed to support product design based on user experience.The proposed approach was validated and implemented in real settings. For consumer products, the time-to-market pressure and market share competition are intensive due to the shortening product life cycles. Product form design that contributes to the user experience (UX) is critical to distinguish the product from others. However, few studies have been done for exploring the relationship between UX and the design of product form. To fill the gaps, this study aims to propose a UNISON framework for data-driven innovation to capture the user experience and preference among the factors of product form designs to derive useful rules. An empirical study was conducted for the product design of wearable devices of a world leading Electronics Manufacturing Service (EMS) company with experimental designs of the subjects with different backgrounds to extract their UX to derive design rules. The results have shown practical viability of the proposed approach to assist the designers to develop product design strategies based on the consumer characteristics and the product UX to launch the preferred products to the corresponding customers to enhance customer satisfaction.
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Nowadays, there are more attentions on cost control and yield enhancement in the semiconductor industry. Many manufacturers have the ability to collect the physical data called Status Variables Identification (SVID) by sensors embedded in... more
Nowadays, there are more attentions on cost control and yield enhancement in the semiconductor industry. Many manufacturers have the ability to collect the physical data called Status Variables Identification (SVID) by sensors embedded in the advanced machines during the manufacturing process. To maintain the competitive advantages, process monitoring and quick response to yield problem are pivotal in detecting the cause of the faults with the help of the sensor data. To state the physical nature of certain SVID, we usually transform SVID into Fault Detection and Classification parameters (FDC parameters) using statistical indicators. The data containing FDC parameters is called FDC data. This study aims to develop a multivariate analysis model to find out the crucial factors which may lead to process excursion among a large amount of FDC data. We proposed a 2-phase multivariate analysis framework: (1) the Least Absolute Shrinkage and Selection Operator (LASSO) is applied for key operation screening. (2) And Random Forest (RF) is used to rank the FDC parameters based on the key operations. Based on the results, domain engineers can quickly take actions responding to low yield problems.
Research Interests:
ABSTRACT Inventory management is challenging and critical problem in the semiconductor memory manufacturing industry. Rapid technological development and short product life cycle cause a high risk of product obsolescence. However,... more
ABSTRACT Inventory management is challenging and critical problem in the semiconductor memory manufacturing industry. Rapid technological development and short product life cycle cause a high risk of product obsolescence. However, manufacturers must still hold a reasonable level of inventory to satisfy the needs of customers when demand is uncertain and lead times are long. In this study, the needs and constraints of this semiconductor manufacturing problem based on inventory days were considered and a linear programming model was constructed with the objective of determining a weighted minimal cost for resolving the problem of multistage inventory management. This study applied the concept of material requirement planning for calculating the inventory information of various stages. In addition, the shortage problem, excessive or inadequate safety stock levels, and capacity balance of technology were considered. Finally, a numerical study was conducted for evaluating the performance of the inventory model, and the results showed that the model can assist decision makers in early preparation of inventories. Under the same condition of demand fulfillment, the proposed model can be used to reduce the unnecessary inventory of downstream banks and avoid the risk of future inventory obsolescence, thereby enhancing competitiveness.
Research Interests:
Research Interests:
Capacity planning in semiconductor manufacturing industry is a challenging task due to its high capital investments, volatile demands, and the long lead time. Current approaches to handle this problem are to model it as an optimization... more
Capacity planning in semiconductor manufacturing industry is a challenging task due to its high capital investments, volatile demands, and the long lead time. Current approaches to handle this problem are to model it as an optimization problem where the uncertain demand is either based on a single forecast, or is decomposed via a finite-scenario structure with an assigned probability for each scenario to reflect the likelihood of occurrence. However, when the uncertainty cannot be decomposed into finite scenarios or when the number of possible scenarios is extremely large, traditional approaches such as the mathematical programming either cannot deal with the problem or may require unreasonable computing time. In this paper, we consider a multiple-period capacity planning problem where the uncertain demand is modeled as a continuous stochastic process over the planning horizon. Moreover, the long lead time and capacity migrations between different products are taken into account to accurately determine the optimal capacity plan. A new framework based on sample path method in simulation optimization is proposed to solve the problem. Comparing to traditional methods, the new framework is much more efficient in terms of the required computing time, as is demonstrated in the computational study.
Research Interests:
Research Interests:
Demand fulfillment and capacity utilization directly affects customer satisfaction, market growth, and the profitability of the company in the semiconductor industry. These characteristics boost the significance of allocating various... more
Demand fulfillment and capacity utilization directly affects customer satisfaction, market growth, and the profitability of the company in the semiconductor industry. These characteristics boost the significance of allocating various customer demands to a number of wafer fabrication facilities (fabs) with different capacity configurations. Before volume production, the introduction of new semiconductor product, namely new tape-out (NTO), requires extremely sophisticated and lengthy qualification with high-cost masks and pilot runs in the qualified fabs. Thus, the NTO allocation will affect future product mix of the qualified fabs, and the flexibility to fulfill the volume demands of the allocated NTOs in the corresponding fabs. This research aims to construct a two-stage stochastic programming (2-SSP) demand fulfillment model. The first stage considers NTO allocation decisions to a number of qualified fabs before the corresponding demand volume is realized. The second stage allocates the capacity to fulfill demand requirements based on the results of four options of capacity reconfiguration: (1) qualifying a product to more than one fab (share); (2) physically transferring a set of masks for a product from one fab to another, where a requalification is required (transfer); (3) moving tools from under-loaded fabs to over-utilized fabs (backup); and (4) utilizing different technologies to capacity inside a fab to support the technology with insufficient capacities (exchange). Both the share and transfer options require long lead time for qualification, whereas the backup and exchange options can be accomplished within a planned timeframe. A numerical study based on real settings is conducted to estimate the validity of the proposed 2-SSP model via values of stochastic solution (VSS) and expected values of perfect information (EVPI). The results showed the benefits of adopting 2-SSP models, especially in an environment with high-demand fluctuation. Furthermore, the proposed 2-SSP can provide near-optimal solutions similar to those of deterministic models with perfect information.
Research Interests:
Research Interests:
Moore's Law indicates that the amount of transistors which can be accommodated on integrated circuit (IC) double every two years in semiconductor manufacturing industry. As a capital intensive and competitive industry, supply chains... more
Moore's Law indicates that the amount of transistors which can be accommodated on integrated circuit (IC) double every two years in semiconductor manufacturing industry. As a capital intensive and competitive industry, supply chains in the semiconductor industry feature high structural complexity and high demand uncertainty. In order to enhance the effectiveness of the enterprise, total resource management (TRM) for semiconductor industry is increasingly crucial. Past studies focused on measuring the effectiveness of either demand fulfillment or capacity utilization, but little researches have done for investigating the inter-coordination on both aspects. This paper proposes overall supply chain effectiveness (OSCE) for the incorporation of demand planning and capacity portfolio based on PDCCCR framework to evaluate the effectiveness of semiconductor supply chain industry. The proposed index is implemented in an empirical study based on a simulation model for semiconductor backend production. Finally, the paper concludes with the discussion of future research directions.
Research Interests:
The chiller machine is one of the most electricity-consuming parts of factory facilities in high-tech industries such as semiconductor and TFT-LCD manufacturing. To reduce variability and optimize the chiller allocation, researchers have... more
The chiller machine is one of the most electricity-consuming parts of factory facilities in high-tech industries such as semiconductor and TFT-LCD manufacturing. To reduce variability and optimize the chiller allocation, researchers have come up with various solutions, but few of them can indeed be widely adopted due to differences of factory layout, machine types, data collections, etc. This study proposes a solution that integrates big data analytics and machine learning techniques to automatically provide recommendations of chiller optimization for energy saving. The optimal chiller adjustment is defined as the condition that the required cooling load for a wafer fab is satisfied while and the electricity consumption is minimized. In the meantime, those adjustment alternatives considering chiller healthy status to obviate inappropriate combinations. Hence, engineers only need to judge the rationality of these recommendations to adjust chillers so that can guarantee operation effe...
Research Interests:
The productivity of bottle neck influences the gross yield of manufacturing system significantly. Since the bottle neck in TFT-LCD array process is photolithography stage, it becomes vital to solve the photo WIP scheduling problem.... more
The productivity of bottle neck influences the gross yield of manufacturing system significantly. Since the bottle neck in TFT-LCD array process is photolithography stage, it becomes vital to solve the photo WIP scheduling problem. Nevertheless, it turns out being much more complex than the usual parallel machine scheduling problem while considering production constraints of photo WIP scheduling problem. Besides dispatching orders to proper machines, auxiliary resource dispatching issues such as mask allocation problem should also be concerned. To empower Industry 3.5 Manufacturing Intelligence for better decision making, the study proposed a modified genetic algorithm encoded with auxiliary capacity planning strategy. Distinguished from other evolutionary algorithm(EA), the proposed Auxiliary Capacity Strategy Genetic Algorithm (ACPS-GA) allows decision makers to resolve photo WIP scheduling problem featuring higher utilization and fast response to the manufacturing system.
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
A dynamic task assignment approach is developed for minimizing the cycle time of business processes.This approach is easier to achieve global optimization and control the resources performing tasks...
Research Interests:
A fab energy efficiency assessment model is established to evaluate the energy saving.An empirical study was conducted to construct OPE indicators for energy conservation.Specific suggestions are provide for each indicator to improve the... more
A fab energy efficiency assessment model is established to evaluate the energy saving.An empirical study was conducted to construct OPE indicators for energy conservation.Specific suggestions are provide for each indicator to improve the energy expenditure. In recent years, global climate change has caused worldwide severe disasters. Countries started to convene CO2 reduction meetings, but they failed to achieve expected outcome. With global greenhouse effect getting worse, governments have more actively made energy saving policies. In 2010, Ministry of Economic Affairs in Taiwan adapted many energy saving strategies, which pushed semiconductor enterprises to take it seriously.Semiconductor industry is both technology and energy intense; therefore, using effective ways to reach energy saving goal has become an important issue. This study aims to establish a Fab energy efficiency assessment model for the semiconductor industry based on SMART decision analysis structure, which used multiple energy saving methods to measure energy cost management efficiency, and considered both theoretical and practical conditions to give suggestions to each indicator to improve the decision making on energy expenditure and distribution. In addition, this study took a Fab in Taiwan as an example to demonstrate the inspection of energy efficiency and construct OPE (Overall Power Energy Effectiveness) indicators, which can help managers to understand the execution of each decision-making unit and properly deploy the resource. The results showed that a Fab must additionally evaluate the energy saving design in expansion and concretely explain energy saving effect and energy management, analyzed the encountered energy difficulties and solutions in 8-inch Fab system, provided a reference for future expansion. The results demonstrated the practical viability of this approach.
Research Interests:
Research Interests:
Research Interests:
Research Interests:
In the era of rapidly increasing notebook market, consumer electronics manufacturers are facing a highly dynamic and competitive environment. In particular, the product appearance is the first part for user to distinguish the product from... more
In the era of rapidly increasing notebook market, consumer electronics manufacturers are facing a highly dynamic and competitive environment. In particular, the product appearance is the first part for user to distinguish the product from the product of other brands. Notebook product should differ in its appearance to engage users and contribute to the user experience (UX). The UX evaluates various product concepts to find the design for user needs; in addition, help the designer to further understand the product appearance preference of different market segment. However, few studies have been done for exploring the relationship between consumer background and the reaction of product appearance. This study aims to propose a data mining framework to capture the user's information and the important relation between product appearance factors. The proposed framework consists of problem definition and structuring, data preparation, rules generation, and results evaluation and interp...
Research Interests:
With advances in information and telecommunication technologies and data-enabled decision making, smart manufacturing can be an essential component of sustainable development. In this era, the promising relevant opportunities to reduce... more
With advances in information and telecommunication technologies and data-enabled decision making, smart manufacturing can be an essential component of sustainable development. In this era, the promising relevant opportunities to reduce costs to boost productivity and improve quality is based on the integration or combination of simulated replicas of actual equipment, Cyber-Physical Systems (CPS) and regionalized or decentralized decision making into a smart factory. However, this integration also presents the industry with different unique challenges. The stream of the data from sensors, robots, and CPS can aid to make the manufacturing smart. Therefore, it would be an increased need for modeling, optimization, and simulation to the value delivery of manufacturing data. This paper aims to outline the approach that was used to develop a system dynamics model to evaluate a superior design of “Industry 4.0” implementation for smart manufacturing.
Research Interests:
The coronavirus pandemic (COVID-19) caused by severe acute respiratory syndrome–coronavirus 2 (SARS-CoV-2) has threatened public health and caused tremendous social and economic losses [...]
Research Interests:
We target the problem of providing 5G network connectivity in rural zones by means of Base Stations (BSs) carried by Unmanned Aerial Vehicles (UAVs). Our goal is to schedule the UAVs missions to: i) limit the amount of energy consumed by... more
We target the problem of providing 5G network connectivity in rural zones by means of Base Stations (BSs) carried by Unmanned Aerial Vehicles (UAVs). Our goal is to schedule the UAVs missions to: i) limit the amount of energy consumed by each UAV, ii) ensure the coverage of selected zones over the territory, ii) decide where and when each UAV has to be recharged in a ground site, iii) deal with the amount of energy provided by Solar Panels (SPs) and batteries installed in each ground site. We then formulate the RURALPLAN optimization problem, a variant of the unsplittable multicommodity flow problem defined on a multiperiod graph. After detailing the objective function and the constraints, we solve RURALPLAN in a realistic scenario. Results show that RURALPLAN is able to outperform a solution ensuring coverage but not considering the energy management of the UAVs.
Research Interests:
As global competition continues to intensity in high-tech industry such as the semiconductor industry, wafer fabs have been placing more importance on the increase of die yield and the reduction of costs. Because of automatic... more
As global competition continues to intensity in high-tech industry such as the semiconductor industry, wafer fabs have been placing more importance on the increase of die yield and the reduction of costs. Because of automatic manufacturing and information integration technologies, a large amount of raw data has been increasingly accumulated from various sources. Mining potentially useful information from such large databases becomes very important for high-tech industry to enhance operational excellence and thus maintain competitive advantages. However, little research has been done on manufacturing data of high-tech industry. Due to the complex fabrication processes, the data integration, system design, and requirement for cooperation among domain experts, IT specialists, and statisticians, the development and deployment of data mining applications is difficult. This chapter aims to describe characteristics of various data mining empirical studies in semiconductor manufacturing, pa...