Improving thermal efficiency and reducing pollutant emissions such as NOx is a critical problem to be solved for conservation and emission reduction in power plant energy. A combustion optimization model was established by quantum genetic algorithm (QGA) optimized least squares support vector machine (LSSVM-QGA). The model predicted boiler thermal efficiency and NOx emissions at average relative error of 0.054% and 1.229%

respectively

which demonstrated high prediction accuracy

generalization ability and applicability. Based on the model

a method of differential evolution algorithm (DE) of self-adaptive scaling/crossover factors and sharing function followed by evolutionary Nash equilibrium was proposed for multi-objective optimization of boiler's combustion. Results show that optimization method based on NASH equilibrium can get optimal set of solutions to operational variables

which improve operating conditions and keep power generation boiler at a stable equilibrium state of combustion.