The Vehicle Routing Problem with Dynamic Occasional Drivers

Abstract

Technological advances, such as smart phones and mobile internet, allow for new and innovative solutions for transportation of goods to customers. We consider a setting where a company not only uses its own fleet of vehicles to deliver products, but may also make use of ordinary people who are already on the road. This may include people who visit the store, who are willing to take a detour on their way home for a small compensation. The availability of these occasional drivers is naturally highly uncertain, and we assume that some stochastic information is known about their appearance. This leads to a stochastic vehicle routing problem, with dynamic appearance of vehicles. The contribution of this paper is a mixed integer programming formulation, and insights into how routes for the company vehicles could be planned in such a setting. The results of the stochastic model are compared with deterministic strategies with reoptimization.

‘Pony express’ crowdsourcing logistics for last-mile delivery in B2C e-commerce: an economic analysis

“…The third B2C crowd local delivery model may be seen as a 'hybrid' solution combining traditional and crowdsourced deliveries. More in detail, an e-commerce player integrates the own van fleet with a group of occasional riders, who still deliver parcels in small quantities, and typically not deviating too much from their original route (Dahle, Andersson, and Christiansen 2017;Macrina et al 2017). This solution has gained the interest of academics, and different contributions may be found in this direction.…”

“…a crowdsourcing initiative involving a large pool of citizens and relying on mobile devices -based on buses and taxis in Singapore and Beijing areas. Some other authors focus instead on the analysis of variables and parameters that could affect the achievement of these cost reductions, such as the uncertainty about the dynamic appearance of drivers during the day (Dahle, Andersson, and Christiansen 2017) or the possibility to set time windows (Macrina et al 2017).…”

“…Two sensitivity analyses were performed in order to test the robustness of the results, and to evaluate the effects on the outcomes of possible changes in the considered scenario, for both the number of riders and the vehicle mix. As a matter of fact, both the availability of riders during the day (Dahle et al, 2017) and the considered transport mean (Kafle, Zou, and Lin, 2017) may strongly affect the performances of a crowdsourcing logistics solution.…”

Recent challenges in Routing and Inventory Routing: E‐commerce and last‐mile delivery

“…A more general setting with respect to the one studied in Archetti et al [14], including customers' time windows and where ODs can serve multiple customers' orders, is studied in Macrina et al [70]. The VRPOD in a dynamic setting has been considered in Arslan et al [17], Dayarian and Savelsbergh [46], and Dahle et al [45]. In particular, Dahle et al [45] study the problem in which ODs reveal their availability in real time.…”

“…The VRPOD in a dynamic setting has been considered in Arslan et al [17], Dayarian and Savelsbergh [46], and Dahle et al [45]. In particular, Dahle et al [45] study the problem in which ODs reveal their availability in real time. Arslan et al [17] and Dayarian and Savelsbergh [46] study the case in which both the customers' orders and the ODs' availability are uncertain and dynamic.…”