The Intermittent Traveling Salesman Problem with Different Temperature Profiles: Greedy or not?

Abstract

In this research, we discuss the intermittent traveling salesman problem (ITSP), which extends the traditional traveling salesman problem (TSP) by imposing temperature restrictions on each node. These additional constraints limit the maximum allowable visit time per node, and result in multiple visits for each node which cannot be serviced in a single visit. We discuss three different temperature increase and decrease functions, namely a linear, a quadratic and an exponential function. To solve the problem, we consider three different solution representations as part of a metaheuristic approach. We argue that in case of similar temperature increase and decrease profiles, it is always beneficial to apply a greedy approach, i.e. to process as much as possible given the current node temperature.