재유입 다중칩 조립라인을 위한 인공신경망 기반의 다중 목적 함수 디스패치 규칙 선택 기법
Published in 한국정보기술학회논문지, 2019
Recently, the semiconductor manufacturers have focused on the production of multiple chip products (MCPs) to achieve high capacities while preserving compactness. As the number of chips to be stacked increases, a re-entrant flow that repeats die attach and wire bonding assembly stages in MCP production is generated, which causes increased flow time and decreased resource utilization. In this paper, we propose a rule selection dispatcher (RSD) based on an artificial neural network, which reduces the flow time and increases the resource utilization. RSD learns the preferences of each dispatching rule according to assembly line data generated by a simulator. Then the proposed dispatcher performs lot dispatching decisions by selecting one of the dispatching rules. The experiments showed that the proposed dispatcher outperformed the existing methods in terms of the performance and computation time.
Recommended citation: 허재석*, 박종헌 (2019), 재유입 다중칩 조립라인을 위한 인공신경망 기반의 다중 목적 함수 디스패치 규칙 선택 기법, 한국정보기술학회논문지, 17(2), 1-11. (KCI)
Download Paper