基于组件式蚁群算法的车辆路径问题研究

Study of vehicle routing problem based on component ant colony algorithm

现有方法仅从蚁群算法的基本结构出发设计软件,缺少软件设计模型的有效指导,很难用来求解不同种类的优化问题。鉴于此,应用组件软件设计方法,提出了一种组件式蚁群算法。该方法力求在结构上直接反映蚁群的本质思想和关键概念,最大程度降低与问题的相关性,强调以接口为中心的设计理念。大量车辆路径问题的结果表明,组件式蚁群算法性能优良,能够有效地求解车辆路径问题。该方法易于理解和使用,具有很强的可重用性和可扩展性,为求解各类优化问题提供了很好的起点和可持续发展的框架。

Existing ant colony algorithm （ACA） software is designed only based on its basic structure. Due to lack the guidance of software design model, it is very difficult to design some practicable ACA software for solving different optimization problems. For this purpose, a component-based ant colony algorithm （CACA） is presented by applying the component-based software engineering thoughtway. This CACA emphasizes the direct correspondence of ACA concepts with its software architecture, the minimal coupling with specific problem structure and the interface-centered design. The experimental results of many vehicle routing problems （VRP） suggest that the CACA is feasible and valid for VRP. This proposed approach is very easy to understand and employ, it has a very robust reusage and expansibility, it indeed provides a good start point and a framework that can continually improve for solving different optimization problems.