Solving optimization problems simultaneously: the variants of the traveling salesman problem with time windows using...

Area of Science:

• Operations Research
• Computer Science
• Optimization Algorithms

Background:

• The Traveling Repairman Problem with Time Windows (TRPTW) and Traveling Salesman Problem with Time Windows (TSPTW) are critical logistics problems with customer-specific delivery time constraints.
• Existing algorithms address TRPTW and TSPTW independently, lacking a unified approach for simultaneous optimization.

Purpose of the Study:

• To develop and evaluate an efficient algorithm that simultaneously solves both TRPTW and TSPTW.
• To leverage knowledge transfer between tasks for improved solution quality in multi-task optimization.

Main Methods:

• A novel algorithm integrating the Multifactorial Evolutionary Algorithm (MFEA) framework with Randomized Variable Neighborhood Search (RVNS).
• The MFEA facilitates knowledge transfer across tasks, while RVNS enhances exploration of the solution space.

Main Results:

• The proposed MFEA-RVNS algorithm demonstrates superior performance compared to the state-of-the-art MFEA in solving TRPTW and TSPTW concurrently.
• The algorithm achieved several new best-known solutions across various datasets, indicating significant improvements in optimization capabilities.

Conclusions:

• The combined MFEA-RVNS approach effectively solves TRPTW and TSPTW simultaneously, outperforming existing methods.
• This integrated strategy offers a promising direction for multi-task optimization problems in logistics and beyond.