Abstract

The rapid growth of urbanization and use of motor vehicles in the recent decades has led to many social and economic problems such as: rising fuel prices, energy crises, environmental problems and traffic congestion. All these problems together have decreased the quality of life of people all around the world. In recent years, municipal planners have increasingly focused on extending policies to promote a culture of using bicycles instead of cars. In many cases, urban planners try to build the infrastructure needed to increase the usage of bicycles and one of the measures that has been widely used by them in recent years is bike sharing programs. In this study, we design a bike sharing network considering the objectives of users and system designers simultaneously. From the customers’ view point, walking short distances before picking up and after dropping off a bike would be a preference and they will be satisfied when they find available bikes or empty docks in the system. From the system designer’s perspective, the objective is to achieve these service levels with the minimum network design cost. To achieve this, we develop a mixed integer linear programming model to minimize the cost of opening stations and transportation costs. We consider the pickup and drop off service level constraints in determining the location, dock capacity and demand allocation to the bike stations. A Mixed Integer Linear Programming model is developed and solved using CPLEX Software. In order to validate the network design solutions, we simulate the results of small to medium size instances in Arena. To solve the larger instances of the problem, a Genetic Algorithm is proposed that uses a heuristic method to generate a part of initial solutions and improves the solutions in its stochastic iterations and reaches near-optimal solutions in a reasonable amount of time. The proposed method is illustrated using the city of Montreal as case study.