Bundle generation for the vehicle routing problem with occasional drivers and time windows

Area of Science:

• Operations Research
• Logistics Management
• Computational Optimization

Background:

• The vehicle routing problem with occasional drivers and time windows (VRP-OD-TW) is a complex logistical challenge.
• Existing VRP-OD-TW models often focus on single requests, limiting flexibility and potential cost savings.

Purpose of the Study:

• To develop and evaluate novel bundling strategies for the VRP-OD-TW.
• To design an auction-based system for assigning customer bundles to occasional drivers efficiently.
• To investigate the impact of spatial-temporal customer representations on bundling effectiveness.

Main Methods:

• An auction-based system was designed for bundle-to-driver assignment.
• Two bundling strategies were proposed utilizing spatial-temporal customer data in 3D space: 3D corridors and 3D clustering.
• Extensive computational experiments were conducted to compare the strategies.

Main Results:

• The 3D corridor bundling strategy significantly outperformed the 3D clustering approach in both solution quality and computational time.
• Both proposed spatial-temporal bundling strategies demonstrated superior performance compared to methods relying solely on spatial information.
• The auction system facilitated attractive bundle offers for drivers and potential cost reductions for the company.

Conclusions:

• Spatial-temporal bundling, particularly using 3D corridors, is a highly effective approach for the VRP-OD-TW.
• The proposed auction mechanism can successfully manage bundle assignments to occasional drivers.
• Integrating temporal dimensions in bundling strategies is crucial for optimizing logistics and achieving cost savings.