PREDICTION OF OPERATOR PERFORMANCE DURING LEARNING OF REPETITIVE TASKS
PREDICTION OF OPERATOR PERFORMANCE DURING LEARNING OF REPETITIVE TASKS
复制标题
DOI：
10.1080/00207547008929848
复制
发表时间：
1970
影响因子：
9.2
通讯作者：
F. Bevis;C. FINNlEAR;D. Towill
中科院分区：
文献类型：
--
作者：
F. Bevis;C. FINNlEAR;D. Towill
在执行重复性任务中的学习效果通常可以用物理系统中常见的指数定律来充分描述。该定律的特征在于上升时间和输出速率的最终值。一种迭代的方法是开发的上升时间和最终值的确定，只使用在学习的早期阶段记录的性能数据，并示出足够准确地预测这些参数用于成本计算和不断更新石灰标准。预测技术的进一步用途包括强调需要加强监督，或更换特定的操作人员。当实验数据是振荡的性质，预测误差大大减少了三点平均调谐到振荡周期。大部分实验数据都是在工业环境中记录的，通常用于长时间的生产运行。
SUMMARY Learning effects in the execution of repetitive tasks may often be adequately described by an exponential law commonly found in physical systems. This law is characterised by the rise time and the final value of output rate. An iterative method is developed for the determination of rise time and final value using only performance data recorded during early stages of learning, and is shown to predict these parameters sufficiently accurately for use in costing and in continuously updating lime standards. Further uses of the predictive technique include highlighting the need for increased supervision, or the replacement of particular operatives. When experimental data is oscillatory in nature, prediction errors are greatly reduced by three-point averaging tuned to the period of oscillation. Much of the experimental data has been recorded in industrial environments, frequently for long production runs.