Pickup and delivery planning for the crowdsourced freight delivery routing problem

Area of Science:

• Operations Research
• Logistics Management
• Computational Optimization

Background:

• Crowdsourced freight delivery relies on solving the Pickup and Delivery Problem (PDP) and Dynamic Vehicle Routing Problem (DVRP).
• Existing research often overlooks the dynamic nature of these problems in real-world crowdsourced scenarios.
• A formal definition of the dynamic DVRP for crowdsourced freight is needed.

Purpose of the Study:

• To define and model the dynamic vehicle routing problem (DVRP) specifically for crowdsourced freight delivery.
• To minimize total service costs, encompassing fixed vehicle, transportation, and delay penalty costs.
• To develop and evaluate efficient algorithms for solving this complex optimization problem.

Main Methods:

• A mixed-integer linear programming model utilizing a rolling-horizon framework was developed.
• An improved partheno genetic algorithm (IPGA) was proposed to solve the combinatorial optimization problem.
• A simulated annealing (SA) algorithm was also implemented for comparative analysis.

Main Results:

• Numerical experiments showed the IPGA outperformed the SA in solving the DVRP.
• The IPGA achieved an average reduction of over 10% in total service costs compared to SA.
• A real-world case study confirmed the practical applicability and effectiveness of the proposed model and algorithms.

Conclusions:

• The developed model and IPGA offer an effective solution for dynamic vehicle routing in crowdsourced delivery.
• The findings provide a practical foundation for implementing optimized routing strategies in real-world logistics.
• Minimizing service costs and ensuring timely delivery are achievable through advanced algorithmic approaches.