基于GPR-TOPSIS方法的充填配比多目标多属性优化

为获得某金矿尾砂胶结充填材料最优配比，基于试验结果，以海水比例、灰砂比和料浆质量浓度为输入参数，以充填体强度、塌落度及泌水率为输出参数，建立了充填配比与其响应量的高斯过程回归模型，分析了不同因素对充填性能的影响程度；采用遗传算法对高斯过程回归模型进行多目标参数优化，获得了Pareto非劣解，在此基础上，引入多属性决策的TOPSIS法对Pareto非劣解进行方案优选，确定了充填最优配比。研究结果表明:高斯过程回归模型相对误差值均小于6%，可靠性高；灰砂比及料浆质量浓度对充填性能影响较为显著，采用海水作为充填水源将降低充填体的强度；经优化后的充填配比与试验结果相符。

Multi objective and multi attribute optimization of filling proportion based on GPR-TOPSIS method

To obtain the optimum proportion of the tailings cemented filling material of a gold mine, through taking the seawater ratio, the cement sand ratio and the mass concentration of slurry as the input parameters, and taking the strength of filling body, the slump and the bleeding rate as the output parameters, a Gaussian process regression model of filling proportion and its responses was established based on the test results, and the influence of different factors on the filling performance was analyzed. The multi objective parameters optimization of the Gaussian process regression model was carried out by using the genetic algorithm, and the Pareto non inferior solutions were obtained. On this basis, the optimal selection of schemes on the Pareto non inferior solutions was conducted by introducing into the TOPSIS method of multi attribute decision making, and the optimum filling proportion was obtained. The results showed that all the relative errors of the Gaussian process regression model were less than 6%, with a high reliability. The cement sand ratio and the mass concentration of slurry had more significant influence on the filling performance, and the strength of filling body reduced when taking the seawater as the filling water source. The filling proportion after optimization was consistent with the test results.