A Queueing Model Based Intelligent Human–Machine Task Allocator

Automatic machines are increasingly being used to help drivers automatically complete tasks; however, the high error rate of automatic machines limits how they might reduce driver task load. Therefore, allocating tasks between human and machine becomes an important question in system design. Existing methods of task allocation do not consider several natural characteristics of human-machine systems simultaneously, including speed-error tradeoff, cognitive modeling of workload, multicriteria decision modeling, dynamic allocation, and global optimum. In this paper, a queueing model-based intelligent task allocator (QM-ITA) that covers the criteria above and optimally allocates tasks between a human operator and an automatic machine is developed. The optimal task allocation algorithm is described in four scenarios that demonstrate how QM-ITA is able to minimize the workload of human operator, minimize system error rate, propose a maximum acceptable error rate of an automatic machine, determine if an automatic machine is necessary for a system, and suggest a maximum acceptable task arrival rate. Further development of the model and the prospects for future research are also discussed.