A mixed closed-open multi-depot routing and scheduling problem for homemade meal delivery incorporating drone and crowd-sourced fleet: A self-adaptive hyper-heuristic approach

Meal delivery services is an enormous competitive market, and the most influential factor in this market is customer experience. As a result, it is crucial to have an efficient delivery system that ensures customer satisfaction regarding on-time, fresh delivery of meals. Accordingly, taking advantage of novel modes of transportation and developing relevant planning approaches can help companies preserve their competitive edge. In this context, the present paper aims to optimize the delivery operations of a homemade meal delivery start-up by adding drones and crowdsourcing, as two innovative modes, to its current system. The addressed problem is an extension of a mixed closed-open pickup and delivery vehicle routing problem. It includes multi -modal transportation fleet and time windows; besides, the meals are time-sensitive and the orders may need to be synchronized. For this purpose, first, a multi-objective mathematical model is devised that considers transportation costs, freshness of the delivered meals, and due-date satisfaction as the objective functions. Afterwards, an efficient self-adaptive hyper-heuristic method is developed to deal with the complexity of the problem. This hyper-heuristic method is based on genetic algorithm and modified particle swarm optimization, and incorporates novel selection and mutation mechanisms. Applying the model to a case study demonstrated that employing drones and crowdsourcing entails 13.7%, 8.5%, and 20.7% improvement in the cost, meal freshness, and weighted due-date satisfaction, respectively.

Automated summary

Meal delivery services is a competitive market, and customer experience is crucial. Enhancing delivery operations by adding drones and crowdsourcing can improve cost, meal freshness, and due-date satisfaction. A mathematical model and an efficient self-adaptive hyper-heuristic method based on genetic algorithm and modified particle swarm optimization are developed.