An Efficient Algorithm for Finding Sets of Optimal Routes

In several important routing contexts it is required to identify a set of routes, each of which optimizes a different criterion. For instance, in the context of vehicle routing, one route would minimize the total distance traveled, while other routes would also consider the total travel time or the total incurred cost, or combinations thereof. In general, providing such a set of diverse routes is obtained by finding optimal routes with respect to different sets of weights on the network edges. This can be simply achieved by consecutively executing a standard shortest path algorithm. However, in the case of a large number of weight sets, this may require an excessively large number of executions of such an algorithm, thus incurring a prohibitively large running time. We indicate that, quite often, the different edge weights reflect different combinations of some "raw" performance metrics (e.g., delay, cost). In such cases, there is an inherent dependency among the different weights of the same edge. This may well result in some similarity among the shortest routes, each of which being optimal with respect to a specific set of weights. In this study, we aim to exploit such similarity in order to improve the performance of the solution scheme. Specifically, we contemplate edge weights that are obtained through different linear combinations of some (``raw'') edge performance metrics. We establish and validate a novel algorithm that efficiently computes a shortest path for each set of edge weights. We demonstrate that, under reasonable assumptions, the algorithm significantly outperforms the standard approach. Similarly to the standard approach, the algorithm iteratively searches for routes, one per set of edge weights; however, instead of executing each iteration independently, it reduces the average running time by skillfully sharing information among the iterations.