Computational Heuristics in Vehicle Communication Infrastructure Optimization

This work studies computational heuristics in the search for urban mobility solutions for the 21st century society. The Delta Network, which is a metric to characterize the performance of vehicular networks, is discussed and two heuristics that perform the allocation of roadside communication units in urban centers (based on the Delta Network) are proposed. The first heuristic operates greedily, making allocations to locations that result in the best momentary performance. The second heuristic uses a genetic algorithm with local search to perform the allocation. The choice of some parameters for the execution of the genetic algorithm, as well as the most promising locations for the deployment of the roadside communication units are studied throughout this work. The proposed heuristics are compared to the intuitive strategy of allocating roadside units at the most popular areas of the road network. The results show that the heuristics achieve better results than the intuitive strategy for all scenarios. The proposed heuristics are also compared with each other, and the results show that the genetic algorithm achieves results slightly superior to the greedy algorithm in the studied scenarios, which indicates that the greedy decisions are well designed to the problem.