Using the Crowd of Taxis to Last Mile Delivery in E-commerce: a Methodological Research

Crowdsourcing is garnering increased attention in freight transport area, mainly applied in internet-based services to city logistics. However, scientific research, especially methodology for application is still rare in the literature. This paper aims to fill this gap and to propose a methodological approach of applying crowdsourcing solution to Last Mile Delivery in E-commerce environment. The proposed solution is based on taxi fleet in city and a transport network composed by road network and customer self-pickup facilities that are 24 hours shops in city, named as TaxiCrowdShipping system. The system relies on a two-phase decision model, first offline taxi trajectory mining and second online package routing and taxi scheduling. Being the first stage of our study, this paper introduces the framework of the system and the decision model development. Some expected results and research perspectives are also discussed. 1 Introduction In E-commerce environment, Last Mile Delivery (hereafter LMD) is the problem of transport planning for delivering goods from e-retailers' hub to the final destination in the area, for example the end consumers' home, see [1] and [2]. Speed and cost are the two crucial success factors to LMD. Faster shipping while with lower cost is the major challenge; nevertheless, it is also a paradox to a certain extend. Indeed, when customers are given a choice between fast and cheap delivery, most of them choose the cheap one, observed by a recent report [3]. The report also infers that that low-cost, speedy two-day delivery corresponds to most customers' expectation, opposite to the one-day delivery policy pursued by giant e-retailers such as Amazon and Alibaba etc. This fact may open up new opportunities to innovative freight transport models [4] for LMD aiming at reducing delivery cost while respecting shipping time,