Robust Multiobjective Optimization for Vehicle Routing Problem With Time Windows.

Area of Science:

• Operations Research
• Optimization Theory
• Transportation Logistics

Background:

• The vehicle routing problem (VRP) with time windows is a critical logistics challenge.
• Real-world VRP scenarios often involve uncertainty in travel times, impacting route efficiency.
• Existing VRP models may not adequately address the dynamic nature of travel times.

Purpose of the Study:

• To develop a robust optimization framework for the VRP with time windows under travel time uncertainty.
• To introduce a novel approach for quantifying and managing uncertainty in VRP.
• To minimize both total distance and the number of vehicles in uncertain environments.

Main Methods:

• A robust multiobjective particle swarm optimization (PSO) algorithm is proposed.
• The algorithm incorporates an advanced encoding/decoding scheme and a robustness measurement metric.
• Two local search strategies (problem-based and route-based) are integrated to enhance solution performance.

Main Results:

• The developed PSO approach effectively handles travel time disturbances.
• The algorithm demonstrates a strong ability to generate robust and near-optimal solutions.
• Experimental comparisons validate the effectiveness of the proposed robust optimization method.

Conclusions:

• The proposed robust multiobjective PSO is a viable method for solving VRP with time windows under uncertainty.
• The approach provides a balance between solution optimality and robustness.
• This research contributes to more reliable and efficient transportation planning in dynamic conditions.