Effects of update frequencies in a dynamic capacitated arc routing problem

The capacitated arc routing problem (CARP) concerns a minimum-cost set of routes for vehicles that provide service on edges in a given graph while ensuring that the total demand in each route does not exceed the vehicle's capacity. This paper concerns a dynamic variant of the CARP. In particular, it focuses on a problem in which new tasks appear over time. We find that simply increasing the number of iterations of a tabu search algorithm does not always lead to a better solution for a dynamic CARP. This paper investigates how the solution quality can be affected by changing the frequency of updating solutions. Furthermore, we investigate whether or not such effect varies with a method of integrating new tasks into the solution at each update.