Research On The Integrated Circuit Back-End Test Scheduling Problems Under Resource Constraints

With the rapid development of information industry, IC foundry has already been the precursor of the national economy. Semiconductor manufacturing is a process of intensive capital and complicated technological process. Then how to improve the utilization of equipment, respond to customers’ requirements quickly, and allocate the limited resources reasonably is the target of semiconductor manufacturing scheduling.The front-end and the back-end constitute the whole production line in semiconductor. Scholars have studied the front-end process a lot and almost obtain the nearly optimal solutions. This makes the issues of back-end problems highlighted. Scheduling the IC chip can improve the benefit of the whole process in the back-end test. The majority of problems in the stage of final text have been proved as NP-hard. Moreover, coupled with the machine capacity constraints and the second resource constraints such as robot hand, try which makes the problem more complicate.This paper researches on the semiconductor manufacturing scheduling problems of the back-end test under secondary resource constraints. First we study the problem of a single batch processor. For this problem, we establish the mathematical model and then we design the variable neighborhood search algorithm (VNS), random key genetic algorithm (RKGA) and batch of insertion algorithm (BIA). Based on the data of simulation test, we compare these algorithms with oven first batch first fit (OFBFF) and the general genetic algorithm. Futher we analyze their advantages and disadvantages. Second we expand the problem of a single batch processing machine to identical parallel batch processing machines. According to two different production processes of parallel batch processing machines, we present two solutions, in which each solution is decomposed into four different stages. Then first batching then machine-batch of insertion algorithm (FBTM-BIA) and first machine then batching-batch of insertion algorithm, (FMTB-BIA) are accordingly proposed to compare the two different solutions. Their effectivenesses are then showed through some experimental data.