Research Interest

Global optimization of the mixed-integer nonlinear programming (MINLP) problems

Global optimization is concerned with the computation of global optimum of a given function in a specified domain. Global optimization of MINLP problems constitutes an active area of research. MINLPs involve finding optimum values of continuous as well as integer decision variables. Use of deterministic branch-and-bound type algorithms that guarantees the global minima of a given MINLPs exists. Several state-of-the-art solvers, such as, AlphaECP, BARON, Bonmin, DICOPT are some of the well-known examples of branch-and-bound type algorithms.

Nuclear reactor modeling and model reduction

Mathematical models of many engineering systems like power systems, chemical process systems and nuclear systems are frequently of high dimension, or possess interacting dynamic phenomena and are nonlinear in nature. These models are often too large for the purpose of control design, analysis and optimization. The cost and complexity of the controller increases as the system order goes high. In some cases the designed controller may not give satisfactory results. The modeling of reactor has been considered. The model of a reactor is nonlinear in nature and has very high dimension. Use of model reduction facilitates the decomposition of the controller structure into separate controller two modules. Thus, it enables simple design along with saving in cost, instead of designing of controller fro whole plant.

Fuzzy logic based control

In practice nonlinearity and complexity dominates the mathematical model of a reactor. Implementation of linear state feedback requires estimation of states such as delayed neutron precursors. Many times in practice it is not possible to get exact system model because of approximate knowledge about the physical system or due to small perturbations in the system parameters. An alternative way to design a control is with model free controller. Fuzzy logic emulates the human knowledge about the control task to be performed rather than technical understanding of a system to be controlled. Fuzzy control is a practical alternative for a variety of challenging control applications since it provides a convenient method for constructing nonlinear controllers via the use of heuristic information. Use of fuzzy controllers for different operating conditions of reactor is tested.