Using Genetic Algorithms to Design Optimal Cognitive Work

Abstract

This paper proposes a method, and associated justification, for the use of genetic algorithms as a way of optimizing the allocation of functions, levels of automation, and cognitive work in system design. Genetic algorithms are useful in finding near-optimal solutions to search problems involving large, combinatorial spaces with multiple objectives. In attempting to determine the best combination of functions and levels of automation in joint cognitive systems, designers typically must consider issues of performance, cost and operator workload at a minimum. Allocation decisions often are left to the subjective judgment of cognitive systems engineers, thus introducing biases on the part of these engineers and uncertainty on the part of the systems engineering community. We anticipate that the use of genetic algorithms will mitigate both of these challenges.