Jayakumar Loganathan's scientific contributions

The parameter estimation optimization with constraints for the nonlinear complex system requires a serious of computation. This paper introduced a novel constrained optimization method named Lagrangian based state transition algorithm (LSTA) to solve problems in distributed cloud computing environment. LSTA with the physical constraints involved in solving the problems which occurs while the conventional techniques are used. In LSTA, the updating of the result to an optimization problem with constraints known as, a state transition. The Lagrangian multiplier used as a constraint for state transition process to estimate the drying process system effectively. The experiments are conducted in the cloud computing environment and simulated results are validated the proposed LSTA methodology for parameter estimation. This method is a promising way for system identification due to its searching competency, enduring performance considering physical limitations and quick convergence.

Spectrum Sensing is playing vital role in Cognitive Radio Networks (CRN). Using cooperative network structure most of the researches had been carried out Channel sensing and proven the better performance than other. In this Energy efficient Sensing, Task assignment and free channel allocation for requested CR’s are mostly varies based on location of CR from Cluster head. Since heterogeneous CR users are participating in the cooperation, their willingness, Battery power, Sensing Accuracy are to be considered for CR’s weight calculation. Otherwise overall sensing results trustworthiness and performance of sensing task may be degraded. Using Multi Criteria Decision Making (MCDM) scheme we have calculated each node weight and given that as input to Multi-level hierarchical clustering technique. Based on the traffic of each blocks, sensing task will be allocated to a cluster. In MCDM, we derived the weight by using combination of MAHP (Modified Analytical Hierarchical Processing) for giving weight for attributes of each CR and VIKOR. Because of considering multiple factors of each CR node, unnecessarily complex sensing task will not be allocated to CR node with lower capability. Similarly, sensing task for channels with predetermined traffic will not be allocated to node with higher capability. Final result shows that increased average lifetime of each participating nodes in the simulation.