A State Aggregation Approach for Stochastic Multiperiod Last-Mile Ride-Sharing Problems

Abstract

The arrangement of last-mile services is playing an increasingly important role in making public transport more accessible. We study the use of ridesharing in satisfying last-mile demands, with the assumption that demands are uncertain and come in batches. The most important contribution of our paper is a two-level MDP framework that is capable of generating a vehicle-dispatching policy for the aforementioned service. We introduce state summarization, representative states, and sample-based cost estimation as major approximation techniques in making our approach scalable. We show that our approach converges and solution quality improves as sample size increases. We also apply our approach to a series of case studies derived from a realworld public transport dataset in Singapore. By examining three distinctive demand profiles, we show that our approach performs best when the distribution is less uniform and the planning area is large. We also demonstrate that a parallel implementation can further improve the performance of our solution approach.

Dynamic and Stochastic Shipment Matching Problem in Multimodal Transportation

“…With the development of computing power, online decisions can be made incorporating dynamic and stochastic information. In the literature, dynamic and stochastic approaches have been investigated in many areas, such as vehicle routing problems (13)(14)(15)(16), pickup and delivery problems (17)(18)(19)(20)(21), resource allocation problems (22)(23)(24), and multimodal container routing and flow control problems (11,12,25). However, to the best of the authors' knowledge, none of the studies in the literature investigated dynamic and stochastic models for the DSSM problem.…”

“…Lowalekar et al ( 20 ) presented a multistage stochastic optimization formulation to consider potential future demand scenarios in online spatial-temporal matching of services to customers and a Benders Decomposition method to deal with large numbers of future scenarios. Agussurja et al ( 21 ) proposed a Markov decision process model for a dynamic ride-sharing problem with stochastic multiperiod demands. Because of the curse of dimensionality, they employed three techniques to speed up the solution process: representative states, state space discretization, and SAA.…”

Data‐Driven Driver Dispatching System with Allocation Constraints and Operational Risk Management for a Ride‐Sharing Platform

“…Spivey and Powell (2004) study a specific version of dynamic vehicle routing, termed as dynamic assignment problem , where each vehicle can serve only one task at a time by modeling the system as a Markov decision process. Agussurja, Cheng, and Lau (2019) similarly build a two‐level Markov decision process framework to study the cases with uncertain demands that come in batches. Wang, Yang, and Zhu (2018) highlight the impacts of appropriate cost‐sharing strategies between drivers and riders, which are commonly centrally determined by ride‐sharing applications according to prescribed rules.…”

Marginalizing Last Mile Logistics Cost through 4th Party Milk Run

“…VRP aims to provide optimal set of routes for a fleet of vehicles to fulfil logistics demands. It has many variants, including VRP with Time Windows [22,23], Capacitated VRP [24,25], green VRP [26,27] and ride-sharing VRP [28,29]. While Milk Run Logistics is essentially a logistics method that uses routing to consolidate the transportation of goods that has the following characteristics: small lot of transport, less than truckload, geographically sparse demands [20], as illustrated in Figure 2.…”