The concept of smart city is frequently used nowadays to refer to the level of integration of cities with citizen services. Mobility and navigation are some examples. When considering disabled users, in a temporary or permanent way, the challenges to provide such services in an efficient way are even bigger. This paper is an extension of a previous work where a mobile application to recommend the most suitable route for citizens was implemented with both Dijkstra and A-Star algorithms. The presented work describes a multi-path approach considering the user might want to choose several destinations in the mobile application and obtain the most suitable route that goes through all of them. Results are presented using the Genetic and the Simulated Annealing Algorithms. Both were evaluated with 100 runs and 2, 4, 8 and 16 destinations. The average algorithm execution time, standard deviation, median value, maximum and minimum execution times were assessed. The Simulated Annealing Algorithm presented better results and always suggested the shorter, or equal, path when comparing to the Genetic Algorithm.