cheapest tour (minimum total distance or some other cost measure associated with the performed trajectory) to sequentially
visit a set of clients (cities, locations) starting and ending at the same client. Vehicles are uncapacitated. The TSP's apparent
simplicity but intrinsic difficulty (NP-hard combinatorial optimization problem) in finding the optimal solution has resulted
in hundreds of publications. A recent literature review can be found in [13]. References to industrial applications and a
comparison of recent approaches can be found in [15].
In the case at hand, nine uncapacitated vehicles are used, a situation known in the literature as them-TSP problem [16,17].
If returning to the depot were not required, the problem would reduce to an Open m-TSPTW, which, to our best knowledge,
has not been studied previously. Open VRP and VRPTW have recently been studied in [32,39], respectively. For a recent
survey on m-TSPs, the reader is referred to [1]. Should the current delivery areas/addresses per route be treated as given,
the problem could be modelled as several separate TSPTWs [21].
All delivery requests are located in the Jyväskylä area, including downtown and rural areas, as demonstrated in Fig. 1.
As the city of Jyväskylä and the subcontractors consider total distance and the number of routes to be the main cost
to drivers in offering the service, the objective function for the computational experiments in Section 4 is defined as the
minimization of the number of vehicles and total distance.