SCHEDULING OPTIMIZATION OF THE TUNISIAN RAILWAY NETWORKS TAKING INTO ACCOUNT PREDICTIVE TASKS

Developments presented in this paper are devoted to the dynalic scheduling of railway transport systems. In this context, our study deals with the implementation of cooperative methodologies to solve railway scheduling problems with predictive or unpredictable demands. Stochastic P-time Petri Nets (SP-TPN) are used for modelling. A new scheduling method based on genetic algorithms and allowing the insertion of predicted jobs is presented, in order to improve a number of traffic criteria such as travel time, total cost and maintenance activities. The aim of the proposed insertion method is to integrate the forecasted demand operations in the availabilities of the rolling stock, in order to minimize the inactivity periods and to increase the traffic rate. Finally, to illustrate the effectiveness and accuracy of the insertion approach, an application to a Sahel Railway networks is outlined.