Abstract
We study a facility location problem motivated by requirements pertaining to the distribution of charging stations for electric vehicles: Place a minimum number of battery charging stations at a subset of nodes of a network, so that battery-powered electric vehicles will be able to move between destinations using “t-spanning” routes, of lengths within a factor t>1 of the length of a shortest path, while having sufficient charging stations along the way. We give constant-factor approximation algorithms for minimizing the number of charging stations, subject to the t-spanning constraint. We study two versions of the problem, one in which the stations are required to support a single ride (to a single destination), and one in which the stations are to support multiple rides through a sequence of destinations, where the destinations are revealed one at a time.
Highlights
Network optimization problems ask us to construct “good” networks subject to various constraints and objectives
In the problem of computing an optimal spanning subgraph, there is a trade-off between the objectives of having a low cost network in terms of the number or weight of the edges and the preservation of shortest path distance in the subgraph, when compared to shortest path distance in the full graph
In this paper we address spanner optimization problems that are motivated by the need for infrastructure in support of electric vehicles
Summary
Network optimization problems ask us to construct “good” networks subject to various constraints and objectives. In the problem of computing an optimal spanning subgraph, there is a trade-off between the objectives of having a low cost network in terms of the number or weight of the edges and the preservation of shortest path distance in the subgraph, when compared to shortest path distance in the full graph. We consider the problem of placing a minimum number of battery charging stations at a subset of nodes of a network, so that battery-powered electric vehicles will be able to move between destinations using routes that are provably close to being shortest paths, while having sufficient charging stations along the way. Our methods can be applied to other metric spaces, such as the L1 distance (measuring “Manhattan” driving distances in a regular grid of streets) or more general road networks
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have