Same-day delivery for online purchases is a recent trend in retail. We introduce multi-vehicle dynamic pickup and problem with time constraints that incorporates key features associated same-day logistics. To make better informed decisions, our solution approach information about future requests into routing decisions. also an analytical result identifies when it beneficial vehicles to wait at the depot. present wide range of computational experiments demonstrate value approach. The results...

In this paper, we analyze how drones can be combined with regular delivery vehicles to improve same‐day performance. To end, present a dynamic vehicle routing problem heterogeneous fleets. Customers order goods over the course of day. These are delivered either by drone or transportation within deadline. Drones faster, but have limited capacity as well require charging after use. context, is not constraint, slow due urban traffic. decide whether an vehicle, policy function approximation...

We consider a stochastic dynamic pickup and delivery problem in which fleet of drivers delivers food from set restaurants to ordering customers. The objective is dynamically control way that avoids delays with respect customers’ deadlines. There are two sources uncertainty the problem. First, customers unknown until they place an order. Second, time at ready restaurant unknown. To address these challenges, we present anticipatory customer assignment (ACA) policy. account for stochasticity...

Operations research requires models that unambiguously define problems and support the generation presentation of solution methodology. In field dynamic routing, capturing joint evolution complex sequential routing decisions stochastic information is challenging, leading to a situation where rigorous methods have outpaced thus making it difficult for researchers engage in science. We provide modeling framework strongly connects application with method leverages rich body route-based planning...

We demonstrate that autonomous-assisted delivery can yield significant improvements relative to today’s system in which a person must park the vehicle before delivering packages. model an autonomous drop off at selected points city where makes deliveries final addresses on foot. Then, picks up and travels next reloading point. In this way, would never need look for parking or walk back place. Based number of customers, driving speed vehicle, walking person, time loading packages, we...

This paper describes a variant of simulated annealing incorporating variable penalty method to solve the traveling-salesman problem with time windows (TSPTW). Augmenting temperature from traditional concept pressure (analogous value multiplier), compressed relaxes time-window constraints by integrating within stochastic search procedure. Computational results validate variable-penalty versus static-penalty approach. Compressed compares favorably benchmark in literature, obtaining best known...

Time-constrained deliveries are one of the fastest growing segments delivery business, and yet there is surprisingly little literature that addresses time constraints in context stochastic customer presence. We begin to fill void by introducing probabilistic traveling salesman problem with deadlines (PTSPD). The PTSPD an extension well-known (PTSP) which, addition presence, customers must also be visited before a known deadline. present two recourse models chance constrained model for PTSPD....

This paper considers a dynamic and stochastic routing problem in which information about customer locations probabilistic future service requests are used to maximize the expected number of customers served by single uncapacitated vehicle. The is modeled as Markov decision process, analytical results on structure optimal policy derived. For case customer, we completely characterize policy. Using results, propose real-time heuristic demonstrate its effectiveness compared with series other...

We develop a family of rollout policies based on fixed routes to obtain dynamic solutions the vehicle routing problem with stochastic demand and duration limits (VRPSDL). In addition traditional one-step policy, we leverage notions pre- post-decision state distinguish two additional variants. tailor our by developing decomposition scheme that achieves high quality large instances reasonable computational effort. Computational experiments demonstrate improve upon performance rolling horizon...

In this paper, we explore same-day delivery routing and particularly how vehicles can better integrate dynamic requests into routes by taking advantage of preemptive depot returns. A return occurs when a vehicle returns to the before delivering all packages currently on-board vehicle. assume that serves in particular area. Beginning day with some known deliveries, seeks serve as well additional new are received throughout day. To requests, must pick up for delivery. contrast previous work on...

Mobile communication technologies enable between dispatchers and drivers hence can fleet management based on real-time information. We assume that such capability exists for a single pickup delivery vehicle we know the likelihood, as function of time, each vehicle's potential customers will make request. then model analyze problem constructing minimum expected total cost route from an origin to destination anticipates responds service requests, if they occur, while is en route. this Markov...

We develop restocking-based rollout policies to make real-time, dynamic routing decisions for the vehicle problem with stochastic demand and duration limits. Leveraging dominance results, we a computationally tractable method estimate value of an optimal restocking policy along fixed route. Embedding our procedure in algorithms, show outperforms priori-based rollout, demonstrating explicitly considering preemptive capacity replenishment approach routing. also demonstrate effectiveness basic...