Neural network approach to design of distributed hard real-time systems

This paper examines the utility of neural networks for optimization problems occuring in the design of distributed hard real-time systems. In other words, it describes how neural networks may also be used to solve some combinatorial optimization problems, such as: computer locations in distributed system, minimalization of overall costs, maximization of system reliability and availability, etc. All requested parameters and constraints in this optimization process fullfil the conditions for design of distributed hard real-time systems. We show that the neural network approach is useful to obtain the good results in the optimization process. Numerical experimentation confirms the appropriateness of this approach.