Abstract
In this paper we apply Estimation of Distribution Algorithms (EDA) to job shop scheduling problems that are known to be NP-hard. EDA is a method that uses a probabilistic model, instead of the crossover and mutation in Genetic Algorithms, to generate individuals in the next generation. We proposed an EDA in which the priority values of operations are estimated and also the relationships between operations of jobs are considered. In the method, the priority value of an operation is estimated with consideration to the relationship between all the former operations. In addition, we examine the method of introducing the correlation coefficient when generating new genes. And also we examine the method using the mixture of EDA and ATC rule that is known to be effective with regard to tardiness. Numerical experiments are used to examine the proposed methods.
