Mixed integer formulations for a routing problem with information collection in wireless networks

Abstract

We study a routing-collecting problem where a system of stations is considered. A vehicle is responsible for collecting information generated continuously in the stations and to deliver it to a base station. The objective is to determine the vehicle route and the collection operations, both physical and wireless, in order to maximize the amount of information collected during a time horizon. Three mixed integer programming models are introduced and a computational study is reported to compare the performance of a solver based on each one of the models.

Heuristics for a vehicle routing problem with information collection in wireless networks

“…In these applications, innovation and research appears most in the development of routing protocols [5,6,19,26], while a smaller set of works addresses the development of vehicle routing strategies, essentially in a two-phase manner [12,14,17,21]. To the better of our knowledge, the works presented in [4,10,15,9] are the only ones to focus on the simultaneous design of a vehicle route and a wireless transmission planning. Notice that the authors in [15] assume a fixed architecture for the vehicle routing (cycle path or zig-zag path).…”

“…In [9], three different objective functions are discussed and experiments are used to investigate how one strategy, i.e., the optimization of one objective function, affects the others and impacts the periodicity of the remaining information. In [10], the exact solution of the WTVRP defined here is investigated. The adoption of the strategy that maximizes the amount of information collected at the end of time horizon allowed the introduction of three different Mixed ILP (MILP) models: one discrete time model and two event based models.…”

“…For works on matheuristics applied to complex routing problems see for instance [1,28]. In the present work, we use the vehicle event model presented in [10] and propose several matheuristic based approaches with the aim of solving large instances of the WTVRP.…”

“…We introduce three heuristics for the construction of an initial solution to the WTVRP. Two of them are matheuristics based on the computational efficient mathematical model introduced in [10], and one is a greedy heuristic. As the size of the MILP model depends on the possible number of visits, the two matheuristics use the MILP model by setting a small number of visits.…”

“…Section 2 states formally the WTVRP while notations and assumptions are presented. In Section 3, we describe the vehicle event model presented in [10] for the WTVRP. Then, in Section 4 the constructive, the improvement heuristics and the strategies combining different types of heuristics are presented.…”