Multi-Core Processor Based Parallel Implementation For Finding Distribution Vectors In Markov Processes

Stochastic processes can be best modeled using Markov processes and these are being used to describe many real time applications. Determining of a distribution / state vector for a Markov process efficiently and swiftly is always be a challenging task. Specially for a high order state and / or for a process having very large number of states. This paper deals in a parallel implementation of finding distribution vectors of a Markov process using repeated multiplications method on multi-core processors. This is done using newly developed parallel programming model for multi-core processors, The (SPCPM)-P-3 at NED University. The results obtained show that the proposed parallel implementation has greater performance and enhanced speedup over the many other serial and parallel implementations for the same process. Besides the implementation found more scalable and easily adoptable for higher number of cores.