Assessing search performance with a simulation model

Abstract

A computer simulation model for visual search is described in general terms. The model has many parameters but the simulation is basically about visual lobe area and search strategy. In the simulation model, the visual lobe area may be defined in different ways and a variety of shapes may be used. Search strategies may be random or systematic with varying amount of lobe overlap possible, or purely random. Using experimental search performance data for 30 subjects with known visual lobe sizes the simulation model produced results close to actual performance. For 20 of the subjects the model showed that they probably used a strategy somewhere between random and random with some lobe overlap, indicating a moderate degree of memory and a rough idea of lobe area. Results for the remaining subjects could be simulated by using a less efficient strategy in terms of lobe overlap. © 1998 John Wiley & Sons, Inc.