How Humans solve Scheduling Problems: Analysis of Human Behavior in the Plan-A-Day Task

by Stefani Nellen

Institution: Universität Heidelberg ; Thes
Year: 0
Record ID: 1112838
Full text PDF: http://archiv.ub.uni-heidelberg.de/volltextserver/2129/


Abstract This thesis explores features of the Plan-A-Day (PAD) task by Funke & Krüger (1993) and presents an analysis of specific aspects of scheduling behavior. The PAD task permits learning at a declarative as well as on a procedural level. Declarative learning in the PAD domain can conceptualized as the accumulation of experience about the feasibility of partial schedules. An 'explorative pattern” is defined which characterizes a scheduling process that implements the strategy of accumulating experience. Procedural learning is hypothesized to take place at the level of the mental arithmetic that is necessary to check schedules in advance (forward checking). It is shown how declarative and procedural learning must work together to enhance scheduling in the PAD task. These assumptions are further investigated in two studies. In the first study, the 'explorative pattern” defined in this thesis is confirmed by the analysis of empirical data. In this study, it also showed that participants who explored little in their first PAD session performed worse in the second session, while the performance of participants who explored much during the first trial improved slightly. Furthermore, in this study a considerable increase of forward checking between the two PAD sessions was found, confirming the assumption about forward checking being the basic method in working with PAD. In the second study, participants evaluated partial schedules. It showed that the feasibility of other appointments is an important reason for evaluating partial schedules. However, according to this study, the actual choice of an appointment is less related to forward checking but to criteria of the appointments. This is interpreted as a qualification of the role of forward checking as the basic skill underlying performance in PAD by implying that there is a preliminary selection process that precedes it.