Comparison of Results of Economic Load Dispatch Using Various Meta-Heuristic Techniques Rajkumar Duraisamy 1* , Gokul Chandrasekaran 1 , Maniraj Perumal 2 , Ramesh Murugesan 2 1 Department of EEE, Velalar College of Engineering and Technology, Affiliated to Anna University, Chennai 600025, India 2 Department of EEE, M. Kumarasamy College of Engineering, Affiliated to Anna University, Chennai 600025, India Corresponding Author Email: rajkumar@velalarengg.ac.in https://doi.org/10.18280/jesa.530217 ABSTRACT Received: 15 November 2019 Accepted: 28 February 2020 The power sector of India is in a huge catastrophe in satisfying the energy requirement of the public due to incessant exhaustion of fossil fuels. The nonstop exhaustion of fossil fuels, rising power needs and increasing production cost of power requires economic operation at the generation side and economic utilization at the consumer side. Economic dispatch is the process of determining the optimal power output from ‘n’ number of generators to meet the demand at low cost subject to certain constraints. Economic dispatch ensures the optimal generation of power at low cost from thermal power plants. The mathematical formulation of economic dispatch problems is usually done by the piecewise quadratic fitness function. This article compares the results generated from various techniques such as Lambda Iteration (LI) method, Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Quantum Particle Swarm Optimization (QPSO) and Shuffled Frog Leaping Approach (SFLA). LI method is a traditional method of solving economic load dispatch which works on the concept of equal incremental cost (λ). GA works on Darwin’s theory of evolution, where the population of individual solutions is modified repeatedly to obtain the optimal solution in the population. PSO is derived from the concept of swarm intelligence, where the best solution is found using the values of personal best and global best in the population. QPSO is basically derived from the PSO. SLFA is obtained from the concept of food- frogs used to find an accurate solution to our power system problem. In this paper, the best fuel cost and execution time was found from QPSO, SFLA compared with LI, GA and PSO methods. These approaches are applied for three and thirteen generator system and the convergence characteristics, heftiness was explored through comparisons from different approaches discussed earlier. The results are hopeful and it suggests that shuffled frog leaping algorithm is very effectual in terms of both the minimized fuel cost obtained and the execution time. Keywords: economic dispatch, particle swarm optimization, quantum behaved swarm intelligence, shuffled frog leaping, Lambda Iteration, fuel cost 1. INTRODUCTION The power utilities all over the world are putting their maximum effort to produce power in a very highly efficient manner to reduce the production cost. The total power generation includes auxiliary power usage of all power generation units. India produced a very high amount of power in 2013 surpassing Russia and Japan with a share of 4.7% globally. During 2014-15, the electricity generation (1010kWh) from utilities as well as non-utilities was higher compared with the electricity consumption of 746 kWh. The consumption of electricity in the agricultural field is very high when compared with all the fields in the year 2014-2015. Even though the electricity tariff in India is very cheaper, the per capita consumption seems quite poor as compared to all the nations [1, 2]. The values of power from all the thermal units can be found by using computer software which should be within the limits of each unit with the demand satisfied. Since the electric power cannot be stored in large amounts, a critical need arises for optimal economic operation of all power plants. Few percent of fuels saved using economic operation result in a lesser production cost in power generation systems. The dissimilar thermal generation units due to the various aspects such as distance, location, and efficiency result in various operating costs. This results in an optimum power generation schedule to decide the capacity of each unit is of very major importance to meet demand at a minimal cost. Also, the power generating cost of all units does not linearly vary with the amount of power it produces. The only way to obtain a profitable schedule is by considering the limits and constraints of the corresponding unit. The foremost aim in power generation is to meet the load demand compensating the power losses which is also a function of power generation. The corresponding improvement in all the unit outputs can result in a significant amount of cost saving [3]. As of now, the energy centers calculate the values of the coordination equations using conventional methods by adjusting the generator values which matches the required load and losses which should result in optimal generation cost. These equations can be easily resolved by interactively changing the load until the sum of the output of the generator matches the load, a failure of the device which should result in a minimum cost of generating electricity at the same time. The country should have an energy policy in such a way that more energy should be produced with minimum cost and losses to serve the feeders. The traditional methods will take more time Journal Européen des Systèmes Automatisés Vol. 53, No. 2, April, 2020, pp. 289-295 Journal homepage: http://iieta.org/journals/jesa 289