基于改进PSO的煤矿安全资源配置问题研究

发布时间：2018-02-02 05:16

本文关键词： 煤矿安全资源配置粒子群算法 Levy飞行机制　出处：《郑州大学》2017年硕士论文　论文类型：学位论文

【摘要】：安全作为煤矿生产和发展贯穿始终的目标,其优劣制约着煤矿安全水平的高低。我国煤炭生产企业为保障良好的安全状态,逐年增加企业安全资源投入,但现实是由高投入提升的安全水平却十分有限。其主要原因是煤炭生产企业内部资源配置存有竞争关系。因此,如何在保障安全水平条件下对煤炭生产企业的安全资源配置进行优化,是企业急需解决的一个重要现实问题。基于煤炭生产企业的现实考虑,在动态时变的生产环境下提高企业的安全水平。本文选用BP神经网络理论及改进PSO方法建立安全资源配置模型并对其进行实证研究。本文的侧重点及创新点包括:(1)构建了煤矿安全资源配置模型。采用BP神经网络训练安全资源投入及安全水平间的非线性关系,以安全水平最高作为资源配置模型的目标函数,根据安全资源投入最小的范围约束,将其作为资源耗费的约束条件,继而建立安全资源配置模型。(2)改进PSO算法用于模型求解。煤炭生产企业的安全资源配置属于多变量、多约束且非线性的函数关系,极易导致粒子群优化算法深陷部分最优且寻优精度低下等状况。对此,本文提出了将Levy飞行机制改进粒子群算法(LF-PSO),将该算法应用至求解安全资源模型,以使煤炭生产企业安全资源配置得到有效优化。(3)以Z煤集团下设矿井作为实证研究对象。为验证安全资源配置模型的可靠且可行,选用Z煤集团进行实证分析,通过BP神经网络方法拟合目标函数的非线性关系,继而建立矿井的资源配置模型,并运用LF-PSO对资源配置模型进行寻优。通过研究表明:煤矿内部关系复杂,常规的PSO算法容易出现“早熟”、进化后期收敛慢等现象。而本文在常规PSO算法基础上,将Levy飞行机制引入PSO算法,提出的基于Levy飞行机制的粒子群算法,能够快速且有效地提高寻优精度和收敛速度。通过LF-PSO计算模型最优解能够达成较为合适的安全资源配置,使企业处于稳步改善的安全水平。因此本文不但丰富了安全资源配置范畴的理论系统,同时还为我国煤炭生产企业配置安全资源提供了科学的决策依据。
[Abstract]:Safety as the coal mine production and development throughout the goal, its advantages and disadvantages restrict the level of coal mine safety. In order to ensure a good state of safety, China's coal production enterprises have increased their investment in safety resources year by year. However, the reality is that the safety level promoted by high investment is very limited. The main reason is that there is a competitive relationship between the internal resource allocation of coal production enterprises. How to optimize the allocation of safety resources in coal production enterprises under the condition of ensuring safety level is an important practical problem that enterprises urgently need to solve, based on the realistic considerations of coal production enterprises. In this paper, BP neural network theory and improved PSO method are used to set up the security resource allocation model and to carry on the empirical research to it. The emphasis and innovation of this paper are put forward. Points include:. (. 1) the model of coal mine safety resource allocation is constructed and the nonlinear relationship between safety resource input and safety level is trained by BP neural network. Taking the highest safety level as the objective function of the resource allocation model, according to the minimum range constraint of the security resource input, it is regarded as the constraint condition of the resource consumption. Then the security resource allocation model. 2) the improved PSO algorithm is used to solve the model. The safety resource allocation of coal production enterprises belongs to the multi-variable, multi-constraint and nonlinear functional relationship. Particle swarm optimization (PSO) algorithm is prone to be trapped in partial optimization and low precision. In this paper, the particle swarm optimization algorithm (PSO) is improved by Levy flight mechanism. The algorithm is applied to solve the security resource model. In order to effectively optimize the allocation of safety resources in coal production enterprises, take the coal mine under Z coal group as the empirical research object, to verify the reliability and feasibility of the safety resource allocation model. Using Z coal group to carry on the empirical analysis, through the BP neural network method to fit the nonlinear relationship of the objective function, and then to establish the mine resource allocation model. LF-PSO is used to optimize the resource allocation model. The study shows that the coal mine internal relations are complex, and the conventional PSO algorithm is easy to appear "precocious". In this paper, based on the conventional PSO algorithm, the Levy flight mechanism is introduced into the PSO algorithm, and the particle swarm optimization algorithm based on the Levy flight mechanism is proposed. The optimization accuracy and convergence rate can be improved quickly and effectively, and the optimal solution of the LF-PSO model can achieve a more appropriate allocation of secure resources. Therefore, this paper not only enriches the theoretical system of safety resource allocation, but also provides scientific decision basis for China's coal production enterprises to allocate safety resources.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD791;F426.21

【相似文献】