Cognitive Models for Learning to Control Dynamic Systems

The public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing the burden, to the Department of Defense, Executive Service Directorate (0704-0188). Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number. Report developed under STTR contract for topic " Cognitive models for learning to control dynamic systems" demonstrated a swarm intelligence learning algorithm and its application in unmanned aerial vehicle (UAV) mission planning. A new UAV assignment model was developed that reduces the dimension of the solution space and is easily adapted by computational intelligence algorithms. A version of particle swarm optimization (PSO) was applied to accomplish the mission optimization. Numerical experimental results illustrate that it efficiently achieves the optima and demonstrates the effectiveness of combining the model and PSO to solve complex UAV assignment problems. The time to complete mission plans for operationally realistic scenarios is reduced by 3-4 orders of magnitude compared with the mixed-integer linear programming approach being used by AFRL at WPAFB. A computer game was also developed to investigate how humans interact with swarm intelligence. The game is based on an NK landscape. It is concluded that the combination of a human-swarm team may have advantages in certain environments, such as dynamic decision making tasks. Dynamic systems are comprised of the following main features: a series of control signals to achieve an optimum, a system that changes as a consequence of the control signals, and other internal/external factors. Thus the control signals are not independent, so that subsequent signals depend on earlier ones. The control signals can either be continuous or be a series of discrete actions. Real world examples include navigational control, battlefield decisions, logistics planning, etc. The goal of analyzing dynamic systems is to develop the ability to characterize all possible trajectories of the system given all possible initial conditions. A state machine is one possible model of a dynamic system. The state machine comprises the set of all possible states of the system, …