Doron Myersdorf

R&D productivity is defined as the evaluated ratio between the outputs with their long-term consequences, and the inputs invested in achieving those outputs. The paper describes the methodology of constructing and implementing a measurement and evaluation system for R&D quality and productivity in high-tech industries. It summarizes continuous fieldwork done with five leading Israeli firms, which focused on establishing and implementing such a system

A discrete event simulation methodology was developed to assist in managing test wafer usage in semiconductor fabs. The purpose of modeling test wafer usage is to predict the number of new test wafers required, test wafer WIP levels, and how to downgrade test wafers to reduce costs of purchasing new test wafers. The test wafer simulation methodology is a detailed yet accurate way to predict test wafer consumption. The methodology has been implemented in a 200 mm development facility, resulting… Show more

A discrete event simulation methodology was developed to assist in managing test wafer usage in semiconductor fabs. The purpose of modeling test wafer usage is to predict the number of new test wafers required, test wafer WIP levels, and how to downgrade test wafers to reduce costs of purchasing new test wafers. The test wafer simulation methodology is a detailed yet accurate way to predict test wafer consumption. The methodology has been implemented in a 200 mm development facility, resulting in considerable cost savings by reducing the overall WIP levels of test wafers. Show less

Object process methodology may be used for the system modeling of research and development (R&D) domain. This is a tool used to assist R&D meet the needs of clients. To assess the performance of R&D, the object-process analysis (OPA) can be utilized. The elements of the system, its environment and their interaction may be examined using OPA. The framework used in the paper enables the user to analyze the issues that R&D managers should take note in the management, planning and assessment of R&D.

Semiconductor manufacturing is one of the most complicated production environments for such reasons as: reentrant product flows; variant batching; clean room environments (imposing constraints on the layout and operations); complicated equipment to fix; and inherent conflict between the need for customized products on the one hand, and high volume, low cost production on the other. Opportunities for productivity improvement, what the authors call “visions of excellence in semiconductor… Show more

Semiconductor manufacturing is one of the most complicated production environments for such reasons as: reentrant product flows; variant batching; clean room environments (imposing constraints on the layout and operations); complicated equipment to fix; and inherent conflict between the need for customized products on the one hand, and high volume, low cost production on the other. Opportunities for productivity improvement, what the authors call “visions of excellence in semiconductor manufacturing” are hindered by these roadblocks. They propose some indicators for the identification of each of the top five roadblocks and the possible industrial engineering solutions for tackling each of them. These top five roadblocks have been identified based on their industry experience and the interviews they have had with many managers in the semiconductor industry. Show less

This paper presents a methodology that provides a vehicle for the selection and ranking of layout design objectives. Moreover, the same methodology can be used to create a decision matrix to evaluate different alternative layouts under analysis. In this case, the methodology combines the ranking of the layout objectives with the relative performance of each layout alternative along each objective in order to recommend the “best” layout alternative. This structured methodology is based on a… Show more

This paper presents a methodology that provides a vehicle for the selection and ranking of layout design objectives. Moreover, the same methodology can be used to create a decision matrix to evaluate different alternative layouts under analysis. In this case, the methodology combines the ranking of the layout objectives with the relative performance of each layout alternative along each objective in order to recommend the “best” layout alternative. This structured methodology is based on a decision making tool known as the Analytical Hierarchy Process (AHP). These concepts are illustrated through several case studies taken from actual layout design projects in the semiconductor manufacturing industry Show less

The issue of cycle time reduction and its impact on a company's competitive edge has been gaining considerable attention recently. Generally speaking, shorter cycle times result in better customer satisfaction, lower work-in-process (WIP), higher yield, and better capacity given tool inventory and facility constraints. This paper provides a brief review of key concepts related to cycle time and describes a methodology for cycle time reduction projects in semiconductor wafer fabrication… Show more

The issue of cycle time reduction and its impact on a company's competitive edge has been gaining considerable attention recently. Generally speaking, shorter cycle times result in better customer satisfaction, lower work-in-process (WIP), higher yield, and better capacity given tool inventory and facility constraints. This paper provides a brief review of key concepts related to cycle time and describes a methodology for cycle time reduction projects in semiconductor wafer fabrication facilities, including the critically important implementation road map step. Finally, a case study is presented to illustrate the effectiveness and potential gains of the proposed cycle time reduction methodology. Show less

This paper discusses the widely known make-or-buy problem. In the semiconductor industry this problem is generally known as the “foundry decision”. This industry has experienced an increased level of wafer fabrication outsourcing in recent years and it is expected that manufacturers will continue to make extended use of foundries in the future. This paper advances the notion that foundry decisions should be assessed along four principal dimensions or objectives: (1) maximization of strategic… Show more

This paper discusses the widely known make-or-buy problem. In the semiconductor industry this problem is generally known as the “foundry decision”. This industry has experienced an increased level of wafer fabrication outsourcing in recent years and it is expected that manufacturers will continue to make extended use of foundries in the future. This paper advances the notion that foundry decisions should be assessed along four principal dimensions or objectives: (1) maximization of strategic competitive performance, (2) maximization of managerial performance, (3) minimization of sourcing risks, and (4) maximization of financial performance. This multiple criteria approach avoids the common pitfall of relying exclusively on financial factors when weighing the pros and cons of each sourcing alternative Show less

This paper describes the concepts underlying a new staffing tool, which evaluates staffing requirements, plan future shift sizes, perform sensitivity analysis, and locate and measure improvement opportunities in terms of worker operations. The model is based on M/M/S/K queuing theory to analyze the operator-machine-lot (OML) interference especially in semiconductor manufacturing processes. Unlike classical approaches, in which the machines are treated as customers, here wafer lots are taken as… Show more

This paper describes the concepts underlying a new staffing tool, which evaluates staffing requirements, plan future shift sizes, perform sensitivity analysis, and locate and measure improvement opportunities in terms of worker operations. The model is based on M/M/S/K queuing theory to analyze the operator-machine-lot (OML) interference especially in semiconductor manufacturing processes. Unlike classical approaches, in which the machines are treated as customers, here wafer lots are taken as the designated customers to be serviced. A comparison of a simulation, OML, and a traditional M/M/S/K model (where the machines are modeled as customers) shows that OML model yields much closer results to the simulation than the traditional M/M/S/K model. Furthermore, from the computational point of view OML is more efficient than the simulation model which can be very time consuming especially in the complex semiconductor manufacturing environment. Thus, OML model is a superior alternative to the classical interference modeling and facilitates staffing recommendations that improve overall equipment effectiveness (GEE) and therefore productivity Show less