A discrete-event driven metaheuristic for dynamic home service routing with synchronised trip sharing

It is common practice in the home service industry that each staff member operates a separate vehicle to visit customers. Facilitating trip sharing and walking policies allows reducing the number of required vehicles, however, often does not succeed due to operational difficulties in routing and planning. In particular, coordinating arrival times of staff members and vehicles at pickup locations introduces major complexity. Previous work in this field focusses on static problem settings where all data is known in advance. Real world-operations, however, are dynamic as cancellations or new requests can happen at any time. This requires decisions to be made in real-time. To assist planners, we propose a flexible discrete-event driven metaheuristic to deal with dynamic routing and scheduling scenarios using combined trip sharing and walking. The computational experiments show that our approach generates solutions in a fast and efficient way, thus, facilitating real-world operations and enabling rescheduling and rerouting. [Received 23 May 2015; Revised 14 October 2015; Accepted 8 December 2015]