Service scheduling and vehicle routing problem to minimise the risk of missing appointments

This research studies a workforce scheduling and vehicle routing problem where technicians drive a vehicle to customer locations to perform service tasks. The service times and travel times are subject to stochastic events. There is an agreed time window for starting each service task. The risk of missing the time window for a task is defined as the probability that the technician assigned to the task arrives at the customer site later than the time window. The problem is to generate a schedule that minimises the maximum of risks and the sum of risks of all the tasks considering the effect of skill levels and task priorities. A new approach is taken to build schedules that minimise the risks of missing appointments as well as the risks of technicians not being able to complete their daily tours on time.

We first analyse the probability distribution of the arrival time to any customer location considering the distributions of activities prior to this arrival. Based on the analysis, an efficient estimation method for calculating the risks is proposed, which is highly accurate and this is verified by comparing the results of the estimation method with a numerical integral method.

We then develop three new workforce scheduling and vehicle routing models that minimise the risks with different considerations such as an identical standard deviation of the duration for all uncertain tasks in the linear risk minimisation model, and task priorities in the priority task risk minimisation model. A simulated annealing algorithm is implemented for solving the models at the start of the day and for re-optimisation during the day.

Computational experiments are carried out to compare the results of the risk minimisation models with those of the traditional travel cost model. The performance is measured using risks and robustness. Simulation is used to compare the numbers of missed appointments and test the effect of re-optimisation.

The results of the experiments demonstrate that the new models significantly reduce the risks and generate schedules with more contingency time allowances. Simulation results also show that re-optimisation reduces the number of missed appointments significantly. The risk calculation methods and risk minimisation algorithm are applied to a real-world problem in the telecommunication sector.