Research On Scheduling Method Of Semiconductor Production Line Based On Symbiotic Organisms Search Algorithm

Semiconductor production line is a representative complex flexible manufacturing system.It has many aporias such as large process flows,reentrant flows and high uncertainty.For this system,it is more important to focus on getting high quality scheduling scheme efficiently with good performance.In this paper,a semiconductor production line is selected as the background.To obtain the high quality scheduling method more quickly,multi-strategy optimization algorithm for single objective problem and efficient multi-objective intelligent scheduling algorithm are studied around the key and aporias in photolithography and etching stages of semiconductor manufacturing system.The contents of the study are as follows:1.To increase the low accuracy for stepper scheduling problem with low yield,we propose an improved Symbiotic Organisms Search algorithm based on multi-strategy fusion.In this method,local search is added to deepen the search depth of Symbiotic Organisms Search algorithm.The search breadth of this algorithm is extended by opposition-based learning.Catastrophic phase helps this algorithm to jump out of local optimal.The original algorithm has few parameters and strong robustness.This section has significantly improved its search results based on retaining its excellent framework.2.To increase the efficiency of multi-objective scheduling in the etching stage,an efficient multi-objective intelligent scheduling algorithm based on multi-objective Symbiotic Organisms Search and entropy-based termination criteria are studied to minimize makespan and total tardiness.This algorithm uses multi-objective Symbiotic Organisms Search as the foundtion.This frame executes algorithm generation by generation.Entropy-based termination criteria determine the convergence epoch of the algorithm.This method saves a lot of time for parameters adjustment and subsequent redundant iterative process.Simulation experiments show that algorithms in our research can effectively solve the problem of long loss time,complex adjustment and huge computational burden in scheduling process.This research has a high value in practical application.