Analytic Cognitive Task Allocation: a decision model for cognitive task allocation

This paper introduces Analytic Cognitive Task Allocation (Analytic CTA) as a decision model for cognitive task allocation. Analytic CTA can support system designers systematically analyse alternative allocation choices, evaluate them through criterion-based quantitative judgements and integrate partial judgements towards a final allocation decision. Analytic CTA was applied in the design of cognitive task allocation by testing three hypotheses. The hypotheses postulated that Analytic CTA can effectively support system designers in three ways: (a) in the identification of task requirements; (b) in the definition of design trade-offs between human allocation and automation; and (c) in the derivation of the allocation decision. Two groups of system designers allocated cognitive functions from three task scenarios in production planning and control. The experimental group, which was supported by the decision model, made allocation decisions which expressed weak preference for either human allocation or automation, consistent with the experimental tasks which had been selected as to elicit weak preferences. The control group relied on the early selection of an automation technology as the means of structuring the design problem. This involved reducing the task to a version suitable for automation, introduced a pro-automation bias and resulted in allocation decisions expressing strong preference for automation. These results were found in support of the experimental hypotheses.