In this seminar, Bu� �ra Çınar will give a talk about Pricing, bundling, and compensatio n decisions in dynamic crowdsourced delivery problems.
Cro wdsourced delivery leverages the unused transport capacity of vehicle s already on the road to perform parcel deliveries. While it can pote ntially tackle challenges in urban delivery, it also introduces uniqu e planning challenges, notably the operator’s lack of direct contro l over driver availability and acceptance behavior. Compensation, now , also is a decision of the operator, which affects drivers’ behavi or. In addition, the operator can offer bundles of tasks, which furth er shape drivers’ responses and increase the problem’s complexity . Therefore, we study a setting in which tasks and drivers arrive dyn amically and stochastically, explicitly model driver- and offer-depen dent acceptance probabilities, and jointly determine bundles and comp ensation. We formulate the problem as a Markov Decision Process and s olve it using Value Function Approximation within an Approximate Dyna mic Programming framework. Preliminary results show clear advantages over benchmark policies.
DESCRIPTION: In this seminar, Buğra Çınar will give a talk about Pr icing, bundling, and compensation decisions in dynamic crowdsourced d elivery problems. Crowdsourced delivery leverages the unused transpor t capacity of vehicles already on the road to perform parcel deliveri es. While it can potentially tackle challenges in urban delivery, it also introduces unique planning challenges, notably the operator’s lack of direct control over driver availability and acceptance behavi or. Compensation, now, also is a decision of the operator, which affe cts drivers’ behavior. In addition, the operator can offer bundles of tasks, which further shape drivers’ responses and increase the p roblem’s complexity. Therefore, we study a setting in which tasks a nd drivers arrive dynamically and stochastically, explicitly model dr iver- and offer-dependent acceptance probabilities, and jointly deter mine bundles and compensation. We formulate the problem as a Markov D ecision Process and solve it using Value Function Approximation withi n an Approximate Dynamic Programming framework. Preliminary results s how clear advantages over benchmark policies. END:VEVENT END:VCALENDAR