Human Interaction with Mission Planning Search Algorithms

摘要: Assigning missions to missiles such as what is needed prior to Tomahawk strikes is an example of a complex resource allocation and optimization problem which includes a large problem space with elements of uncertainty. Attempting to achieve an “optimal” solution is problematic in the military due to high risk environments, dynamic constraints, and the need for rapid decisions under time pressure, which may or may not be improved as a result of the addition of an automated algorithm. In this study we report on the development of mission-missile planning decision support software, and how humans interact and collaborate with a heuristic search algorithm in the context of Tomahawk missionmissile planning. Of particular interest is how the human operator generates mission-missile assignments under different levels of automation: 1) One with low levels of automation (eg, sorting and filtering assistance), 2) Interactive levels of automation in which the human and algorithm work together to solve a problem, and 3) A mission-planner that makes all assignments and the human must determine whether or not the automated solution is correct. Both human and the heuristic search algorithm performance will be discussed as well as the impact of different levels of automation on situation awareness and the implications for operational use.