Presented in this research paper is an attempt to apply a cat swarm optimization (CSO)-based algorithm to the urban transit routing problem (UTRP). Using the proposed algorithm, we can attain feasible and efficient (near) optimal route sets for public transportation networks. It is, to our knowledge, the first time that cat swarm optimization (CSO)-based algorithm is applied to cope with this specific problem. The algorithm’s efficiency and excellent performance are demonstrated by conducting experiments with both real-world as well as artificial data. These specific data have also been used as test instances by other researchers in their publications. Computational results reveal that the proposed cat swarm optimization (CSO)-based algorithm exhibits better performance, using the same evaluation criteria, compared to most of the other existing approaches applied to the same test instances. The differences of the proposed algorithm in comparison with other published approaches lie in its main process, which is a modification of the classic cat swarm optimization (CSO) algorithm applied to solve the urban transit routing problem. This modification in addition to a variation of the initialization process, as well as the enrichment of the algorithm with a process of improving the final solution, constitute the innovations of this contribution. The UTRP is studied from both passenger and provider sides of interest, and the algorithm is applied in both cases according to necessary modifications.
In this contribution a comparative study of modern heuristics on the school timetabling problem is presented. More precisely, we investigate the application of two population-based algorithms, namely a Particle Swarm Optimization (PSO) and an Artificial Fish Swarm (AFS), on the high school timetabling problem. In order to demonstrate their efficiency and performance, experiments with real-world input data have been performed. Both algorithms proposed manage to create feasible and efficient high school timetables, thus fulfilling adequately the timetabling needs of the respective high schools. Computational results demonstrate that both algorithms manage to reach efficient solutions, most of the times better than existing approaches applied to the same school timetabling input instances using the same evaluation criteria.
This paper deals with the school timetabling problem. The problem was formulated as encountered in a typical Greek high school. A local version of the particle swarm optimization algorithm was developed and applied to the problem at hand. Results on well-established benchmark instances showed that the proposed algorithm achieved the proven optima provided from an integer programming method presented in earlier research. In almost all cases, the current algorithm beat the integer programming method, either concerning the lower bound yielded or the execution time needed.