基于量子遗传模糊推理系统的煤与瓦斯突出预测模型  

Prediction model of coal and gas outburst based on quantum genetic fuzzy inference system

作　　者：郭金栋[1] GUO Jindong(School of Energy Engineering,Huainan Vocational and Technical College,Huainan,232001,China)

机构地区：[1]淮南职业技术学院能源工程学院,安徽淮南232001

出　　处：《华北科技学院学报》2023年第6期30-37,共8页Journal of North China Institute of Science and Technology

基　　金：安徽省高校自然科学研究重点项目(2022AH052998);安徽省高等学校省级质量工程项目(2022zygzsj052);淮南职业技术学院自然科学研究项目(HKJ22-2)。

摘　　要：为提高煤与瓦斯突出危险程度预测的准确性,提出一种基于自适应神经模糊推理系统(ANFIS)结合改进实数编码量子遗传算法(IRQGA)的预测模型IRQGA-ANFIS。用基于数据驱动的方法从样本数据直接提取模糊规则,建立煤与瓦斯突出ANFIS预测模型。针对ANFIS预测准确率较低以及模糊推理系统参数量大的特点,采用IRQGA对模糊推理系统进行训练。IRQGA引入秃鹰算法的阿基米德螺线空间搜索机制更新个体;用差分变异策略更新种群最差个体,保持种群多样性;用高斯-柯西变异策略扰动优秀个体使其快速脱离局部极值区,加快算法收敛速度。实验结果表明,IRQGA在高维复杂问题优化中比实验对比算法具有更好的优化性能;IRQGA-ANFIS模型的预测准确率达94.44%;所建模型30次独立运行的MAE均值相较对比模型分别降低了0.0245和0.1184,MSE均值分别降低了0.0162和0.1849,RMSE均值分别降低了0.0172和0.1721。IRQGA-ANFIS具有更高的预测准确率和更好的预测能力。A prediction method based on adaptive neuro-fuzzy inference system and improved real-coded quantum genetic algorithm is proposed to improve accuracy of coal and gas outburst dangerous level.The fuzzy rules of ANFIS is extracted directly from sample data by data-driven method,and an adaptive neuro-fuzzy inference system for coal and gas outburst prediction is established.In view of the low accuracy of ANFIS prediction and the large number of parameters of fuzzy reasoning system,an improved quantum genetic algorithm is used to train adaptive neuro-fuzzy inference system.The Archimedes spiral space search mechanism of the bald eagle search algorithm is introduced into real-coded quantum genetic algorithm to update individual,and the differential mutation mechanism is mutate the worst individual to maintain the diversity of the population,and Gauss-Cauchy mutation is used to alternate the optimal individuals to help them escape from the local extremum region quickly and accelerate the iteration speed of the algorithm.The different forecast methods are conducted on typical engineering practical data of coal and gas outburst.The results show that IRQGA can yield a superior optimization performance to other algorithm,and the prediction accuracy of the IRQGA-ANFIS method is 94.44%,and the MAEs in 30 rounds of independent operation of the model built in this paper reduce by 0.0245,0.1184 on average,and the MSEs reduce by 0.0162,0.1849 on average,and the RMSEs reduce by 0.0172,0.1721.The model built in this paper has better forecast ability and more accuracy for predicting the outburst fatalness.

关 键 词：煤与瓦斯突出 预测 ANFIS 实数编码量子遗传算法 阿基米德螺线空间搜索 高斯-柯西变异 

分 类 号：TD713[矿业工程—矿井通风与安全] X936[环境科学与工程—安全科学]