基于遗传蚁群强化学习算法的露天矿运输路径优化试验研究  

Experimental Study on Optimization of Transportation Path in Open Pit Mines Based on Genetic Ant Colony Reinforcement Learning Algorithm

作　　者：苏树鹏[1] SU Shupeng(School of Artificial Intelligence Technology,Guangxi Technological College of Machinery and Electricity,Nanning,Guangxi 530007,China)

机构地区：[1]广西机电职业技术学院人工智能技术学院,广西南宁530007

出　　处：《矿业研究与开发》2024年第12期221-227,共7页Mining Research and Development

基　　金：广西高校中青年教师科研基础能力提升项目(2022KY1076);广西职业教育教学改革研究项目省部级重点项目(GXGZJG2022A011)。

摘　　要：针对复杂的矿山地形进行路径规划,容易陷入局部最优解的问题,融合遗传蚁群算法和强化学习算法的优点,提出了一种基于遗传蚁群强化学习算法的矿山运输优化模型,实现矿山运输路径的高效优化。首先利用遗传算法对解空间进行高效搜索,生成一组初始的运输路径以探索更广阔的解空间;其次引入蚁群算法,在搜索过程中发现最优解即最优运输路径;最后引入强化学习算法,通过模拟智能体的决策过程并根据环境反馈调整运输路径,使模型逐渐学习到更优的决策策略,在不断迭代的过程中逐步提高运输效率。经过试验验证,相较于传统遗传蚁群算法,遗传蚁群强化学习算法求得的最短运输距离缩短了20.06%,运输成本降低了12.60%,寻找最优路径的时间减少了51.55%。该优化模型能够有效地提高运输效率、降低运输成本,为矿山运输提供高效和环保的解决方案。A mining transportation optimization model based on genetic ant colony reinforcement learning algorithm is proposed to address the problem of easily getting stuck in local optimal solutions when planning paths for complex mining terrain. This model combines the advantages of genetic ant colony algorithm and reinforcement learning algorithm to achieve efficient optimization of mining transportation paths. Firstly, using the genetic algorithm to efficiently search the solution space and generate an initial set of transportation paths to explore a broader solution space. Secondly, introducing the ant colony algorithm to discover the optimal solution during the search process, which is the optimal transportation path. Finally, the reinforcement learning algorithm was introduced to simulate the decision-making process of intelligent agents and adjust transportation paths based on environmental feedback, so that the model can learn better decision-making strategies and can gradually improve transportation efficiency through continuous iteration. Experimental verification shows that compared to traditional genetic ant colony algorithm, the shortest transportation distance obtained by the genetic ant colony reinforcement learning algorithm has been reduced by 20.06%, transportation cost has been reduced by 12.60%, and the time to find the optimal path has been reduced by 51.55%. This optimal model can improve transportation efficiency, reduce transportation cost, and provide efficient and environmentally friendly solutions for mining transportation.

关 键 词：露天矿运输 遗传蚁群强化学习算法 路径规划 运输成本 

分 类 号：TD57[矿业工程—矿山机电]