基于人工鱼群算法的能源配送问题的研究

发布时间：2018-01-07 18:05

本文关键词：基于人工鱼群算法的能源配送问题的研究　出处：《内蒙古农业大学》2015年硕士论文　论文类型：学位论文

【摘要】：物质配送在物流活动过程中充当着十分重要的角色,是与消费者直接相关联的重要环节。然而,配送成本在物流总成本中占着很高的比重。如何采用更科学、合理的配送方法逐渐成为学者们关注的问题,在努力缩减运输成本、缩短配送时间等方面做出了大量的探索和研究,且能源配送问题是物流配送问题中的一种。基于此,各种智能优化算法纷纷涌现,如遗传算法、蚁群算法、人工蜂群算法、模拟退火算法、人工鱼群算法等。这些算法的出现为能源配送问题的解决提供了新的思路、方法和工具,但是能源配送问题本身比较复杂和特殊,属于NP-hard(非确定性多项式)问题。仅仅用基本的智能算法来求解往往找不到满意的最优解,所以这些基本的仿生智能算法需要不断的改进和创新。论文利用改进的人工鱼群算法来求解某煤炭公司的能源配送问题,由于人工鱼群算法本身的局限性,如前期收敛过快导致难以搜寻到最优区域,后期收敛不稳定导致结果不够精准等。针对这些不足之处,论文在认真研究基本人工鱼群算法的基础上,对该算法做了两个方面的改进和创新。首先,引入虚拟人工鱼的概念对基本人工鱼算法鱼群的行为做了改进,用以扩大人工鱼群的搜寻范围,提高算法在前期对最优区域全局搜索的精度。其次,引入衰减因子并且与模拟退火算法相结合,使得该算法后期的局部寻优结果更优。此外又将模拟退火算法作了改进,引入二次交换的思想,以充分发挥混合算法的优势来弥补人工鱼群算法后期收敛精度不高的不足；同时将人工鱼群算法与遗传算法相结合,提出了一种新的混合算法。最后将各混合算法的寻优性能进行了仿真比较。能源配送问题比较复杂,为使得寻优结果更优,所以论文将问题分为两步来解决。第一步,装车方案的选择;第二步,配送路径的寻找,并对其分别进行建模。最后编码后进行仿真实验,将得出的结果和基本算法加以比较。通过对结果的对比、分析,得出改进后的人工鱼群算法和改进的混合人工鱼群算法在解决能源配送问题上比基本人工鱼群算法更具优势。
[Abstract]:The material distribution plays an important role in the process of logistics activities, is directly related to the consumers of the important part. However, the distribution cost accounted for a high proportion in the total cost of logistics. How to adopt more scientific and reasonable distribution method has gradually become the scholars concern, in an effort to reduce transportation costs, make a lot of research and exploration to shorten the delivery time, and the distribution of energy is a problem in logistics distribution. Based on this, a variety of intelligent optimization algorithms have emerged, such as genetic algorithm, ant colony algorithm, artificial bee colony algorithm, simulated annealing algorithm, artificial fish swarm algorithm. The algorithm provides a new way of thinking in order to solve the problem of energy distribution, methods and tools, but the energy distribution is complex and special, which belongs to the NP-hard (non deterministic polynomial) problem. Only with the basic intelligent algorithm The optimal solution can not find a satisfactory solution, so these basic bionic intelligence algorithm requires continuous improvement and innovation. The energy distribution problem of a coal company, the improved artificial fish swarm algorithm used, due to the limitations of artificial fish swarm algorithm, such as pre convergence too fast cause is difficult to find the optimal area. Convergence leads to unstable results is not accurate enough. Aiming at these shortcomings, based on the careful study of the basic artificial fish swarm algorithm, the algorithm is improved and innovation in two aspects. Firstly, introducing the concept of virtual human artificial fish behavior of artificial fish swarm algorithm basic improvements were made in the search in order to expand the artificial fish swarm algorithm to search the optimal region, improve the overall accuracy in the early stage. Secondly, introduces the attenuation factor and combined with the simulated annealing algorithm, the algorithm of late The local search results better. In addition, the simulated annealing algorithm is improved, the introduction of two exchange ideas, in order to give full play to the advantages of the hybrid algorithm to compensate for the lack of artificial fish swarm algorithm convergence precision is not high; at the same time, by combining the artificial fish swarm algorithm and genetic algorithm, this paper proposes a novel hybrid algorithm the hybrid algorithm. The optimization performance is simulated. The distribution of energy is more complex, so that the optimization result is better, so the paper will be divided into two steps to solve the problem. The first step, the loading scheme selection; the second step, with the search route, and the two models. Finally, after encoding the simulation experiment, the results of the basic algorithm and comparison. By comparison, the results of the analysis, we conclude that the improved artificial fish swarm algorithm and improved hybrid artificial fish swarm algorithm to solve the energy distribution. The problem is more advantageous than the basic artificial fish swarm algorithm.

【学位授予单位】：内蒙古农业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP18

【参考文献】