An estimation of distribution algorithm coupled with the generalized Mallows distribution for a school bus routing problem with bus stop selection

Although the estimation of distribution algorithms were originally designed for solving integer or real-valued domains, this contribution applies the algorithms mentioned to deal with a permutation-based problem, called school bus routing problem with bus stop selection, using the generalized Mallows distribution as an attempt to describe and obtain an explicit probability distribution over a set of school bus routes. In addition, a mutation operator is considered for improving the estimation of the central permutation, a parameter of the Mallows distribution. Different and diverse instances served as input and test parameters in order to show that permutation-based optimization problems such as the school bus routing problem with bus stop selection can be solved by means of a probability model, and improving the estimation of the central permutation helps the performance of the algorithm.