Abstract. This paper considers a stochastic location-allocation problem for a capacitated bike sharing system (S-L&A-CBSS), in which bike demand is uncertain. To tackle this uncertainty, a Sample Average Approximation (SAA) method is used. Because this problem is an NP-hard problem, a hybrid greedy/evolutionary algorithm based on Genetic Algorithm (GA) and Particle Swarm Optimization (PSO), namely greedy GA-PSO, is embedded in the SAA method in order to solve the given large-sized problems. The performance of the proposed hybrid algorithm is tested by a number of numerical examples and used for empirical test based on Tehran business zone. Furthermore, the associated results show its e ciency in comparison to an exact solution method in solving small-sized problems. Finally, the conclusion is provided.

In this paper, the bike-sharing problem was extended by considering safety in addition to system costs. Moreover, we determined the cost and safety levels for different kinds of stations. For solving the problem of conflicting objective functions, we used the NSGA-II and MOPSO algorithms and compared them. The results confirmed that the NSGA-II algorithm performs better than MOPSO for considering different solutions to the bikesharing system with safety design problem. In the second stage, a multi-objective model was transformed to a linear single-objective model to find a preferred solution. A genetic algorithm (GA) was developed to solve the proposed large-scale bike-sharing model, and the results were compared with the solution obtained by commercial software. The results showed that the proposed GA outperforms the commercial software solution approach in large-scale instances.

No abstract