Abstract

AbstractAttended home delivery (AHD) is a popular type of home delivery for which companies typically offer delivery time slots. The costs for offering time slots are often double compared to standard home delivery services (Yrjölä, 2001). To influence customers to choose a time slot that results in fewer travel costs, companies often give incentives (discounts) or penalties (delivery charges) depending on the costs of a time slot. The main focus of this paper is on determining the costs of a time slot and adjusting time slot pricing accordingly, i.e., dynamic pricing. We compare two time slot cost approximation methods, a cheapest insertion formula and a method employing random forests with a limited set of features. Our results show that time slot incentives have added value for practice. In a hypothetical situation where customers are infinitely sensitive to incentives, we can plan \(6\%\) more customers and decrease the per-customer travel costs by \(11\%\). Furthermore, we show that our method works especially well when customer locations are heavily clustered or when the area of operation is sparsely populated. For a realistic case of a European e-grocery retailer, we show that we can save approximately \(6\%\) in per-customer travel costs, and plan approximately \(1\%\) more customers when using our time slot incentive policy.KeywordsTime slot managementDynamic pricingVehicle routingMachine learningCost approximation

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call