Schedule Nervousness Reduction in Transport Re-Planning

This article reports about investigations on a transport scheduling scenario. The generated schedule requires a repeated revision in order to incorporate additionally released requests. Neither the release time nor the location of the associated customer site can be forecasted. Special attention is paid to keep the generated schedules stable, which means that once made scheduling decisions should be maintained (announced arrival times at customer sites, etc.) Initially, several measures for the nervousness degree are proposed. In computational simulations it turns out that a very high nervousness degree appears if the same schedule update strategies are applied for different workloads. As a remedy, the adaptation of the used schedule update strategy to the current schedule performance is proposed. A prototypic algorithm framework is presented and assessed is comprehensive numerical simulation experiments. The adaptation of the schedule update strategies contributes to a significant reduction of some observed nervousness degrees. It reinforces the preservation of once fixed arrival times at customer sites.