Tackling the Crowdsourced Shared-Trip Delivery Problem at Scale with a Novel Decomposition Heuristic
Authors:
Dingtong Yang,
Michael F. Hyland,
R. Jayakrishnan
Abstract:
This paper presents a set-partitioning formulation and a novel decomposition heuristic (D-H) solution algorithm to solve large-scale instances of the urban crowdsourced shared-trip delivery (CSD) problem. The CSD problem involves dedicated vehicles (DVs) and shared personal vehicles (SPVs) fulfilling delivery orders, wherein the SPVs have their own trip origins and destinations. The D-H begins by…
▽ More
This paper presents a set-partitioning formulation and a novel decomposition heuristic (D-H) solution algorithm to solve large-scale instances of the urban crowdsourced shared-trip delivery (CSD) problem. The CSD problem involves dedicated vehicles (DVs) and shared personal vehicles (SPVs) fulfilling delivery orders, wherein the SPVs have their own trip origins and destinations. The D-H begins by assigning as many package delivery orders (PDOs) to SPVs as possible, where the D-H enumerates the set of routes each SPV can feasibly traverse and then solves a PDO-SPV-route assignment problem. For PDO-DV assignment and DV routing, the D-H solves a multi-vehicle routing problem with time-window, tour duration, and capacity constraints using an insertion heuristic. Finally, the D-H seeks potential solution improvements by switching PDOs between SPV and DV routes through a simulated annealing (SA)-inspired procedure. The D-H outperforms a commercial solver in terms of computational efficiency while obtaining near-optimal solutions for small problem instances. The SA-inspired switching procedure outperforms a large neighborhood search algorithm regarding run time, and the two are comparable regarding solution quality. Finally, the paper uses the D-H to analyze the impact of several relevant factors on city-scale CSD system performance, namely the number of participating SPVs and the maximum willingness to detour of SPVs. Consistent with the existing literature, we find that CSD can substantially reduce delivery costs. However, we find that CSD can increase vehicle miles traveled. Our findings provide meaningful insights for logistics practitioners, while the algorithms illustrate promise for large real-world systems.
△ Less
Submitted 17 June, 2024; v1 submitted 10 March, 2022;
originally announced March 2022.
Dynamic Modeling and Real-time Management of a System of EV Fast-charging Stations
Authors:
Dingtong Yang,
Navjyoth J. S. Sarma,
Michael Hyland,
R. Jayakrishnan
Abstract:
Demand for electric vehicles (EVs), and thus EV charging, has steadily increased over the last decade. However, there is limited fast-charging infrastructure in most parts of the world to support EV travel, especially long-distance trips. The goal of this study is to develop a stochastic dynamic simulation modeling framework of a regional system of EV fast-charging stations for real-time managemen…
▽ More
Demand for electric vehicles (EVs), and thus EV charging, has steadily increased over the last decade. However, there is limited fast-charging infrastructure in most parts of the world to support EV travel, especially long-distance trips. The goal of this study is to develop a stochastic dynamic simulation modeling framework of a regional system of EV fast-charging stations for real-time management and strategic planning (i.e., capacity allocation) purposes. To model EV user behavior, specifically fast-charging station choices, the framework incorporates a multinomial logit station choice model that considers charging prices, expected wait times, and detour distances. To capture the dynamics of supply and demand at each fast-charging station, the framework incorporates a multi-server queueing model in the simulation. The study assumes that multiple fast-charging stations are managed by a single entity and that the demand for these stations are interrelated. To manage the system of stations, the study proposes and tests dynamic demand-responsive price adjustment (DDRPA) schemes based on station queue lengths. The study applies the modeling framework to a system of EV fast-charging stations in Southern California. The results indicate that DDRPA strategies are an effective mechanism to balance charging demand across fast-charging stations. Specifically, compared to the no DDRPA scheme case, the quadratic DDRPA scheme reduces average wait time by 26%, increases charging station revenue (and user costs) by 5.8%, while, most importantly, increasing social welfare by 2.7% in the base scenario. Moreover, the study also illustrates that the modeling framework can evaluate the allocation of EV fast-charging station capacity, to identify stations that require additional chargers and areas that would benefit from additional fast-charging stations.
△ Less
Submitted 16 December, 2020;
originally announced December 2020.
