Williams B A
Department of Psychology, University of California, San Diego, La Jolla 92093-0109.
J Exp Psychol Anim Behav Process. 1992 Jul;18(3):287-97.
Rats were trained on a discrete-trial version of a concurrent variable-interval (VI) variable-ratio (VR) schedule in which the relative reinforcement rates were varied across conditions. Subjects with the shorter intertrial interval (ITI) had a significant bias toward the VR alternative, as predicted by optimality theory, and were also more likely to choose the VI alternative with longer times since responses to the VI alternative, as predicted by momentary-maximizing theory. Subjects with the longer ITI failed to show either of these effects. Approximation to the matching law was greater with the longer ITI. Thus, matching is not derivative of the processes postulated by optimality or momentary-maximizing theory but instead is in competition with those processes.
大鼠在并发可变间隔（VI）可变比率（VR）强化程序的离散试验版本上接受训练，其中相对强化率在不同条件下变化。如最优性理论所预测，具有较短试验间隔（ITI）的受试者对VR选项有显著偏好，并且如瞬间最大化理论所预测，自对VI选项做出反应后，他们也更有可能选择时间更长的VI选项。具有较长ITI的受试者未表现出这两种效应中的任何一种。ITI较长时对匹配法则的近似程度更高。因此，匹配并非最优性或瞬间最大化理论所假定过程的衍生物，而是与这些过程相互竞争。
