产品研发中面向用户意象的交互式遗传算法评价模式

摘要

针对产品研发过程中了解用户意象的需求，采用交互式遗传算法对产品设计方案实施优化。由于评价噪声影响用户意象的准确获取，对交互式遗传算法的用户评价噪声来源及其与评价精度之间的关系作了分析，对四种不同评价精度下的收敛速度、评价耗时和用户疲劳等指标进行了实验研究。结合产品优化设计的具体案例，论证了低精度评价方法具有降低疲劳、缩短评价时间、提高种群规模等优势，并且可以有效利用数理统计方法进行适应值预测和用户意向分析。%Aiming at the requirement of user image in product development process, Interactive Genetic Algorithms （IGA） was applied to optimize the product design scheme. Owing to the user image＇s acquirement was affected by e- valuation noise, the user evaluation noise source of IGA and its relationship with the evaluation precision were ana- lyzed. Experiments were taken to study the index of convergence speed, time cost, user fatigue features and so on under four different user evaluation precisions of IGA. Combined with actual cases, low-precision evaluation method was demonstrated had advantages in reducing fatigue, time cost and enlarging group scale. In addition, the mathematics stastics method was used to carry out fitness forecasting and user image analysis.