Artificial neural network-based solutions of the interprocess communication problems

Interprocess Communication directly effects process management and operation performanceof the operating system. The performance for the current operating systems can be increased by faster and more efficient process management and scheduling algorithms. In this thesis, Artificial Neural Network-based solution methods areproposed for classical Interprocess Communication problems. Proposed solution is applied to the dining philosophers, readers and writers, producer-consumer, and sleeping barber problems. And, these proposed solutions are compared with semaphore-based solutions. Finally, significant performance improvement is acquired in the dining philosophers and readers and writers problem.