A Combination of Genetic Algorithm and Particle Swarm Optimization for Vehicle Routing Problem with Time Windows

Area of Science:

• Operations Research
• Computer Science
• Artificial Intelligence

Background:

• Vehicle Routing Problems with Time Windows (VRPTW) are complex combinatorial optimization challenges.
• Existing algorithms often struggle with computational burden and premature convergence.
• Efficiently solving VRPTW is crucial for logistics and transportation management.

Purpose of the Study:

• To propose a novel hybrid algorithm combining Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) for VRPTW.
• To improve the efficiency and solution quality for VRPTW compared to existing methods.
• To address limitations such as computational burden and local optima in VRPTW solvers.

Main Methods:

• A hybrid approach integrating GA and PSO is developed.
• Particle real number encoding is employed to decode routes and reduce computation.
• A linear decreasing function balances global and local exploration.
• GA's crossover operator is integrated to prevent premature convergence and local minima.

Main Results:

• The proposed hybrid algorithm demonstrates superior efficiency and competitiveness.
• It achieves more optimal solutions for the VRPTW benchmark problem.
• A new well-known solution for the benchmark problem is presented.

Conclusions:

• The hybrid GA-PSO algorithm is an effective method for solving VRPTW.
• The algorithm overcomes limitations of traditional approaches, offering better performance.
• This research contributes a more efficient and robust solution for complex routing problems.