Multi-swarm PSO and its application in operational optimization of ethylene cracking furnace

The exiting PSO algorithms are analyzed deeply, a multi-swarm PSO (MSPSO) is studied. The whole swarm is divided into three sub-swarms randomly, the first particle group obeys the standard PSO principle to search the optimal result, the second searches randomly inner neighborhood of the optimal result, the third does not care about the optimal result but flies freely according to themselves velocities and positions. So the algorithm enhances its global searching space, enriches particlespsila diversity in order to let particles jump out local optimization points. Testing and comparing results with standard PSO and linearly decreasing weight PSO by several widely used benchmark functions show optimization performance of the algorithm is better. Furthermore, the proposed algorithm is employed to resolve the operational optimization problems of ethylene cracking furnace. The operational optimization results for built cracking model are effective and satisfying.