面向采煤机概念设计的模型推理算法研究

Algorithm Research of Model Reasoning for Shearer Conceptual Design

为充分利用企业经验知识,减少对领域专家的过分依赖,实现数据驱动的采煤机概念设计。面向采煤机概念设计过程,基于ε-支持向量回归理论,利用遗传算法优选模型参数,寻找采煤机条件属性与决策属性间的映射关系,建立了采煤机概念设计推理模型。以UCI标准测试数据集中Mpg、Housing回归测试数据验证了模型具有较高的准确性和可行性。通过该模型设计人员可以由采煤机采高、截深等原始输入参数推理出牵引功率等总体技术输出参数,为传统经验设计提供了科学依据,节约了设计时间与成本。

To make full use of experience knowledge of enterprises, reduces excessive rely on domain experts, and realizes the conceptual design of shearer driven by data. It established the reasoning model for shearer conceptual design based on ε- support vector regression theory, using genetic algorithm to optimize parameters and look for the mapping relationship between condition attributes and decision attributes. It validated the model＇s accuracy and feasibility by regression-test data Mpg and Housing in UCI standard test data set. The model can help designers get the technical parameters such as the traction power by inputting mining height, cutting depth and other primitive parameters, providing a scientific basis for traditional experience design, and saving the design time and cost.