Designing robust routes for demand-responsive transport systems

• An operative context common in many DRT systems is modeled.
• A novel two-stage stochastic model for the design of robust demand responsive transportation systems is proposed.
• A tailored heuristic exploiting the problem structure is presented.
• The value of using stochastic information is assessed.

In this study, we propose an innovative concept for robust demand-responsive transportation (DRT) systems where vehicles may deviate from the planned route to accept late requests, which are unknown during the planning stage. We propose a new formulation of the problem as a stochastic mixed integer program and describe an efficient heuristic procedure that embeds a tabu search approach in a scheme for merging different scenario solutions. The computational results demonstrate the validity of the heuristic and provide useful managerial insights into DRT systems, thereby showing the value of incorporating uncertainty into the planning process.