Stochastic model of effectiveness for man-hardware-software system
This study presents analytical models that provide a realistic measurement of effectiveness for the system involving hardware, software, and human operator. It shows a way to properly integrate the human operator into system effectiveness. System effectiveness is defined as the combined measure of availability and reliability at each task-arrival time. At each task-arrival time, transient human operator behavior is considered in conjunction with the man-hardware-software system state. Since factors such as fatigue, stress, and learning will affect the operator over time, the human performance variables are assumed to vary with time during the mission. The models are developed for both single-machine and N-machine problems and are extended to handle: (1) systems with multiple operating mode, (2) system with several type of task, and (3) system with general failure and repair distribution. For the first problem, the operating levels are assumed to affect the task performance. In the second problem, each type of task is characterized by the performance level of the human operator representing the degree of accomplishment of a specified task. The last extension is considered to handle a more general system by removing the assumption of the Markovian process for the system state.