- A three-stage evolutionary strategy framework based on local search and random reconstruction strategies is proposed to solve VRPTW, aiming to balance exploration and exploitation based on the characteristics of different LNS operator combinations.
- A local search strategy combining random vehicle correlation-based customer removal and regret insertion operators is introduced to further explore the dominant individuals in the population and enhance the quality of individuals.
- A random reconstruction strategy combining random removal and random insertion operators is proposed to help stagnated individuals escape local optima and enhance population diversity.
