Abstract: Aiming at the correlation problem of overall and local comfort in sitting comfort evaluation, a comfort subjective evaluation model based on Gene Expression Programming(GEP)was proposed. The input variable were six local comfort indexes and the output variable were overall comfort score and rating. Experimental data were collected by using comfort subjective evaluation rating scale from subjective perceptions of 48 occupants sitting on a trial chair. Two coupling models were established respectively through GEP comfort classification algorithm, and the prediction effect of two models was compared. The results showed the feasibility of employing GEP in predicting subjective sitting comfort. A subjective evaluation software system of sitting comfort was developed to demonstrate the viability of model in practical application.

Key words: sitting comfort, gene expression programming, ergonomics, subjective evaluation

摘要： 针对坐姿舒适性测评中总体与局部舒适性关联性问题，提出基于基因表达式编程的舒适性主观测评模型。模型的输入变量为6个局部舒适性指标，输出变量为总体舒适性分值与等级。48名被试在实验椅上分别进行坐姿与坐卧姿舒适性主观测评，采用舒适性主观评价量表获取实验数据。通过基因表达式编程舒适性分级算法分别构建了2个耦合模型，并比较了两者的预测效果。实验结果表明，采用基因表达式编程算法预测坐姿主观舒适性是有效的。最后开发了一个坐姿舒适性的主观测评系统，并在实际应用中验证了模型的有效性。

关键词: 坐姿舒适性, 基因表达式编程, 人机工程学