Networks has just published a new special issue on New network models and approaches for logistics and transportation (78:3). The special issue was guest edited by Daniela Ambrosino and Massimo Paolucci. A portion of the editorial by Ambrosino and Paolucci is include below with the full issue available to read here.
Modern logistics and related novel applications require approaches to monitoring and cost reduction that are innovative, highly effective, and conducive to social and environmental sustainability. For this purpose, this special issue collects seven papers proposing innovative network-based models and approaches. This group includes work that adopts both theoretical and applied approaches. Six of these papers share the common modeling framework of transportation networks. Specially, five papers deal with routing problems, considering different application scenarios, one engages in a theoretic analysis in the field of transportation network cooperative games, and one proposes an innovative operation-time–space network that can be easily adapted to model different logistic problems involving time and space decisions.
The paper by Triki et al. considers a public transportation network. In particular, it assumes a ridesharing transportation system that is inspired by the real-life context of a university in the Gulf region. The authors introduce a mathematical model that combines the vehicle routing problem (VRP) with combinatorial auctions. Then, they propose two variants of a new hybrid heuristic algorithm, referred to as the combinatorial auction ridesharing solution framework, which includes three meta-heuristic algorithms—particle swarm optimization, a dragonfly algorithm, and an imperialist competitive algorithm. The authors report results for a set of randomly generated instances, showing the ability of the proposed algorithm to find high-quality solutions with small optimality gaps.
Cerrone et al. propose urban transportation networks that address a green VRP with heterogeneous fleets. The authors determine minimum cost routes for delivering orders placed through e-channels in urban areas, while also taking environmental cost components into account. They propose a mathematical model and a matheuristic for solving this problem in Milan, Italy, one of the smartest cities in Europe and also a municipality that is making big investments to promote the use of bicycles.
Transportation networks are also considered in the work of Passacantando et al. The authors make reference to a game-theoretic analysis of Braess’ paradox, a classical result in the theory of congestion games, which motivates their investigation of why adding a resource (e.g., an arc) to a network may worsen the overall network performance. In particular, the paper considers the occurrence of Braess’ paradox in cooperative games with transferable utility on networks. Then, the Shapley value is evaluated with respect to various network scenarios. The aim is to identify situations in which this value is negative for an arc or node, thus, revealing the negative impact of its insertion into the network. The authors validate their analysis by reporting a cooperative version of Braess’ original example as a case study.
Networks are the modeling framework for the paper by Cassettari et al., who take on a multi-objective VRP associated with a practical application: the logistics operations related to the activation and deactivation of customers’ gas meters. This is an operational planning problem that must be solved daily, and it involves hard time windows with overlaps and maximum working time constraints. The authors present a mathematical formulation and propose a four-step heuristic approach, the effectiveness of which is shown by an experimental campaign on both a real case study and on randomly generated instances.
Networks for modeling routing in transportation can also be found in the flying sidekick traveling salesman problem faced by Dell’Amico et al. Here, a set of electric flying drones are the sidekicks of a truck that delivers parcels, and the solution must take into account a number of particular constraints on the drones’ operations. The authors present novel formulations and valid inequalities for the considered problem, and they perform a comparative study to analyze the impact of several variants of the problem, showing the possible advantages of using multiple drones for delivery operations.
Computer networks are considered in the paper by Addis et al., which deals with the management of service requests in cloud computing facilities. According to the network function virtualization paradigm, part of the computing capacity of network nodes can be reserved to provide network functions in a reliable and cheap way. The problem then lies in both locating the virtual functions in the network nodes and routing the related requests to them. The authors address this problem by proposing four different matheuristics, which they compare on a set of instances generated from the SNDlib 1.0 data-set, varying the size and topology of the network and the virtual network functions’ capacity.
An innovative operation-time–space network is introduced by Ambrosino and Asta. This proposed network is able to deal with different types of capacity and time constraints that characterize most logistics problems. The authors explain how to construct this network and how to model different types of time and capacity constraints. They present a detailed application of this network to schedule port rail shunting operations, and they briefly introduce other applications for testing the proposed network, including the well-known quay crane scheduling problem.