Decision-making of online rescheduling procedures using neuroevolution of augmenting topologies

Abstract

Online scheduling requires appropriate timing of rescheduling procedures, as well as the determination of relevant horizon length. Optimal choices of these quantities are highly dependent on the uncertainty of the scheduling environment and may vary over time. We propose an approach where a neural network is trained to make online decisions on these quantities, as well as on the choice of the rescheduling method (mathematical programming or metaheuristics). In our approach, the neural network is trained using neuroevolution of augmenting topologies (NEAT) in a simulated environment. In this paper, we also optimize the rescheduling interval and horizon length of a conventional periodically occurring rescheduling on a dynamic routing problem. The resulting approach is the baseline for the development of the proposed neural network approach.