A MCM Interconnect Test Generation Optimization Scheme Based on Ant Algorithm and Genetic Algorithm

Abstract

The paper presents a hybrid optimization scheme of ant algorithm (AA) and genetic algorithm (GA) for the interconnect test generation problem in multi-chip module (MCM). In this scheme, the AA is employed to generate the initial candidate vectors for the MCM interconnect test generation, where the pheromone updating rule and state transition rule of AA is designed. Then the GA evolves the candidate vectors generated by AA, using a fault simulator to evaluate the fitness of each candidate vector. Various GA parameters are investigated, including selection operator, crossover operator, crossover and mutation rate, as well as number of generation and population size. The international standard MCM circuit was used to verify the scheme. The results indicate that the performance of the scheme in execution time and fault coverage is comparable to other deterministic algorithms