Cooperative self-configuring nature-inspired algorithm for a scheduling problem

Abstract

One of the crucial challenges related to operational manufacturing planning is an optimal plan search in the current situation using a workflow model. The problem solving is greatly hindered by the rapid growth of the search space with an increase in dimension or the so-called combinatorial explosion. This paper uses two different approaches to solving a hierarchical scheduling problem based on different solution representations. The first approach assumes a search of an optimal project order and then solving of resource-constrained project scheduling problem (RCPSP) for each of the projects using a model based on the greedy principle. In the second approach we are searching for priorities of all actions of all projects and then use them in the process of building a schedule if there are any conflicts when choosing the next action. To solve the problem with both approaches, the paper considers some nature-inspired algorithms such as the intelligent water drops algorithm (IWDs), a genetic algorithm (GA) and ant colony optimization (ACO) as well as a self-configuring version of the last two. The paper shows the efficiency of the application of the coevolution algorithm using IWDs, self-configuring GA and ACO.