Solving Generalized Vehicle Routing Problem With Occasional Drivers via Evolutionary Multitasking.

Area of Science:

• Operations Research
• Logistics Management
• Computational Intelligence

Background:

• The rise of the sharing economy and crowdshipping necessitates new logistics models.
• Existing vehicle routing problems (VRP) do not fully account for occasional drivers or vehicle heterogeneity.
• The vehicle routing problem with occasional drivers (VRPOD) is a recent development addressing these gaps.

Purpose of the Study:

• To introduce a generalized variant, the vehicle routing problem with heterogeneous capacity, time window, and occasional driver (VRPHTO).
• To develop an evolutionary multitasking algorithm (EMA) for simultaneously optimizing multiple VRPHTO instances, inspired by cloud computing's parallel processing needs.

Main Methods:

• A novel evolutionary multitasking algorithm (EMA) was designed to handle multiple VRPHTO tasks concurrently using a single population.
• 56 new VRPHTO instances were created using established vehicle routing benchmarks.
• Empirical studies compared the EMA's performance against a state-of-the-art single-task solver.

Main Results:

• The inclusion of occasional drivers demonstrably reduces overall routing costs.
• The proposed EMA effectively solves multiple VRPHTO problems simultaneously.
• EMA achieved superior optimization performance compared to the single-task solver due to inter-task knowledge transfer.

Conclusions:

• Occasional drivers offer a significant cost-saving potential in goods delivery.
• The EMA provides an effective and efficient approach for multitasking in complex optimization problems.
• Knowledge transfer in multitasking evolutionary algorithms enhances optimization outcomes.