Missing the wood for the wrong trees: On the difficulty of defining the complexity of Complex Problem Solving scenarios

Beckmann, J.F.; Goode, N.

Abstract

In this paper we discuss how the lack of a common framework in Complex Problem Solving (CPS) creates a major hindrance to a productive integration of findings and insights gained in its 40+-year history of research. We propose a framework that anchors complexity within the tri-dimensional variable space of Person, Task and Situation. Complexity is determined by the number of information cues that need to be processed in parallel. What constitutes an information cue is dependent on the kind of task, the system or CPS scenario used and the task environment (i.e., situation) in which the task is performed. Difficulty is conceptualised as a person’s subjective reflection of complexity. Using an existing data set of N = 294 university students’ problem solving performances, we test the assumption derived from this framework that particular system features such as numbers of variables (NoV) or numbers of relationships (NoR) are inappropriate indicators of complexity. We do so by contrasting control performance across four systems that differ in these attributes. Results suggest that for controlling systems (task) with semantically neutral embedment (situation), the maximum number of dependencies any of the output variables has is a promising indicator of this task’s complexity.

Citation

Beckmann, J., & Goode, N. (2017). Missing the wood for the wrong trees: On the difficulty of defining the complexity of Complex Problem Solving scenarios. Journal of Intelligence, 5(2), Article 15. https://doi.org/10.3390/jintelligence5020015

• Journal Article Type: Article
• Acceptance Date: Apr 10, 2017
• Online Publication Date: Apr 13, 2017
• Publication Date: Apr 13, 2017
• Deposit Date: Apr 19, 2017
• Publicly Available Date: Apr 20, 2017
• Journal: Journal of Intelligence
• Electronic ISSN: 2079-3200
• Publisher: MDPI
• Peer Reviewed: Peer Reviewed
• Volume: 5
• Issue: 2
• Article Number: 15
• DOI: https://doi.org/10.3390/jintelligence5020015
• Public URL: https://durham-repository.worktribe.com/output/1381044