For large-scale super-structure optimization on non-convex and non-linear heat exchanger network

conventional intelligent optimization algorithms have poor efficiency and easily fall into local optima. Adaptive competitive swarm optimization algorithm was proposed to optimize no-split stream heat exchanger networks with targeted total annual cost. This method improved abilities of both global and local optimization by attenuated learning of average particle positions and self-adaptive adjusting on weight average of speed. Simulation on two typical cases showed that the proposed algorithm sharply reduced cycles of model being used

shortened optimization time and achieved better optimization results in comparison with quantum particle swarm algorithm.