Decision-based genetic algorithms for solving multi-period project scheduling with dynamically experienced workforce

Abstract

The importance of the flexibility of resources increased rapidly with the turbulent changes in the industrial context, to meet the customers’ requirements. Among all resources, the most important and considered as the hardest to manage are human resources, in reasons of availability and/or conventions. In this article, we present an approach to solve project scheduling with multi-period human resources allocation taking into account two flexibility levers. The first is the annual hours and working time regulation, and the second is the actors’ multi-skills. The productivity of each operator was considered as dynamic, developing or degrading depending on the prior allocation decisions. The solving approach mainly uses decision-based genetic algorithms, in which, chromosomes don’t represent directly the problem solution; they simply present three decisions: tasks’ priorities for execution, actors’ priorities for carrying out these tasks, and finally the priority of working time strategy that can be considered during the specified working period. Also the principle of critical skill was taken into account. Based on these decisions and during a serial scheduling generating scheme, one can in a sequential manner introduce the project scheduling and the corresponding workforce allocations.