Combined Heat and Power Dispatch by a Boost Particle Swarm Optimization

Abstract

Currently, particle swarm optimization (PSO) and its modifications are effectively being used in solving complex optimization problems, simply because of its benefits like the few parameters to control, its implementation is easy, and it has fast convergence. However, PSO has a few drawbacks, like it suffers from premature convergence and/or sometimes convergence speed is slow, and diversity maintenance. The main achievement of PSO relies on the updating velocity of the particle. Algorithm velocity has an important dependency over coefficients multiplied, like acceleration coefficients and inertia weight. This paper introduces an Improved PSO (IPSO) algorithm, to strengthen the capability of PSO and get solutions to complex optimization problems with extra efficiency. In IPSO new inertia weight (progressively decreased) and acceleration factors (steadily increased and/or decreased) are familiarized. Discussed IPSO has been verified on two typical benchmark functions (specifically Rosenbrock and Rastrigrin) then applied for solving combined heat and power economic dispatch (CHPED) problem which contains 4 power, 1 heat and 2 cogenerations-only units. The main objective of the CHPED problem is to minimize the total cost of fuel and emissions though meeting inequality and equality limits plus load demand. The results obtained by experiments show that IPSO is much better than the state-of-the-art algorithms statistically and significantly. Also, the accuracy of convergence and convergence speed of IPSO is increased efficiently.