Using Micro-Architectures of Cognition to Model Macro-Cognitive Systems: A Warfighting Video Game Example

摘要

The use of micro-cognitive architectures to model macro-cognitive systems is motivated by the open ended question as to whether the theories of cognitive science pitched at the level of individual mental processes can be successfully linked to, or bridged with, macro-cognitive phenomena involving multi-agent activity in complex sociotechnical systems. Previous research attempts to use the GOMS human information processor model in sociotechnical systems (West & Nagy, 2007 for references) have shown that GOMS is ill suited to model macro-cognition. SGOMS, or sociotechnical GOMS (West & Pronovost, 2009, West & Somers, 2011 for recent developments) was created as an extension to GOMS in order to cope with such shortcomings, mainly in dealing with interruptions and task switching, two critical issues in sociotechnical systems modeling. The question motivating the present research is whether SGOMS can implement complex, multi-agent, strategic (cooperative and competitive) activities as-is, or requires theoretical and methodological extensions to cope with such interactions. More precisely, what are the implications of cognitive processing constraints on hierarchical command structures? Should the design of command structures consider cognitive constraints, and are there ways of making command structures better? We present an implementation of SGOMS in the ACT-R cognitive architecture, extending SGOMS to strategic (in this context, tactical) activities situated in a virtual environment. SGOMS models were developed, based on task analyses of the massively multi-player online roleplaying video game World of Warcraft, focusing on hierarchical command structures, using two varieties of tactical planning processes, the MDMP (Military Decision-Making Process) and the RPM (Recognition Planning Model).