Hopfield’s Artificial Neural Networks In Multiobjective Optimization Problems of Resource Allocations Control

Some robots cooperating each other to perform connected operations can be considered as a system. This system should be controlled to take advantages of different robots assigning to several operations. Indeed, a control of robot-operation allocations as a sequence of many static optimization task of resource allocations with different input parameters can be formulated. If some robot control problems are transformed to this resource allocation problem, then it is possible to use the proposed below methods. In this paper, analog Hopfield’s artificial neural networks are used by genetic algorithms for solving NP-hard binary multiobjective optimization problems, which can be considered in modeling of resource allocations control. This problem can be solved for improving the efficiency of a few connected robots dining their activities. Moreover, another neural approach for dynamic optimal control is elaborated. Finally, an example of two-layer feed-forward network in the adaptive control system of the underwater vehicle motion is submitted.