Designing a system for a process parameter determined through modified PSO and fuzzy neural network

Abstract

In the manufacturing industry, the key to retaining a competitive advantage lies in increased yield and reduced a number of reworks. Determining the optimal parameters for the process so that the quality characteristics can meet the target is an important strategy. Traditional statistical techniques such as response surface methodology and analysis of variance, whose basic assumptions must be met, are generally used in this regard. In recent years, artificial intelligence has reached a sufficient level of maturity and is extensively being used in various domains. This paper proposes a system based on the modified particle swarm optimizer (PSO) and the adaptive network-based fuzzy inference system (ANFIS) to determine the process parameters. A perturbed strategy is incorporated into the modified PSO. The application of this system is then demonstrated with the determining of parameters in the wire bonding process in the IC packaging industry. Moreover, the performance of the modified PSO is evaluated with testing functions. The results show that the modified PSO yielded a superior performance to traditional PSO. In the optimization of the process parameter, the modified PSO is able to find the optimal solution in the ANFIS model. © 2008 Springer-Verlag Berlin Heidelberg.