| Summary: | 博士 === 大同大學 === 資訊工程學系(所) === 107 === The multiple traveling salesman problem (mTSP) is a combinatorial optimization problem and is an extension of the famous Traveling Salesman Problem (TSP). Not only does the mTSP possess academic research value, but its application is extensive. For example, the vehicle routing problem (VRP) and operations scheduling can all be reduced to mTSP solutions. The mTSP is an NP-hard problem, and multifaceted discussions of its solutions are worthwhile.
This study assigned ants to teams with mission-oriented approaches to enhance ant colony optimization algorithms. Missions were appointed to ant teams before they departed (each ant had a different focal search direction). In addition to attempting to complete its own mission, each ant used the Max-Min strategy to work together to optimize the solution.
The goal of the appointing missions is to reduce the total distance,whereas the goal of using the max–min search method for paths was to achieve Min-Max, or the goal of labor balance. During the solving process, each ant will refer to the pheromone concentration on the paths and the mission tips as their action guidelines. After each round, the mission configuration will be changed in accordance with the state of solution obtained from each mission-coordinated team and the pheromone concentration on the path will be reconfigured. Four main elements were involved in the search process of the ant teams: mission pheromone, path pheromone, greedy factor, and Max-Min ant firing scheme. The experimental results revealed this novel approach to be innovative and effective.
|
|---|
