Research For HSAEA Algorithm And Its Application For Process Modeling

Based on the basic Alopex, which is a single-point evolution algorithm, an AEA (Alopex-based evolutionary algorithm) algorithm was proposed by combining the idea of swarm intelligence evolution algorithm, which makes the variable walk more flexibly and stochastically.AEA algorithm performs well in optimization quality and speed, however it still has some disadvantages. In one hand, the set method of step length is not reasonable. Here, the paper proposed that the moving step length in AEA should be varied with and different kind of search, to make sure at the early stage of search, the algorithm tends to stochastic search in a large scale, and tends to deterministic search around the global optima in the end of the search process, but also has the probability to climb out the local minima. In another hand, the advantage of Harmony search algorithm is introduced to enhance the communication of different individual, and extend the swarm diversity. Then, HSAEA and AEA algorithm are both testing with a series of low-dimension and high-dimension benchmark functions. The results show that HSAEA gains a better performance than AEA, not only possesses a fast convergence speed, but also obtains relatively high solution quality. Later, HSAEA algorithm was applied to model the depth of Ethylene Cracking for the ethylene cracking furnace; simulation results demonstrate that the soft sensor modeling well reflects the actual operation situation of the production facilities, and effectively extracts the interrelation between operational variables of the cracking furnace and cracking depth. Finally, HSAEA was used for the parameter estimation for a heavy oil thermal cracking three lumps model, the reaction kinetics model obtained by our method compares with others reported in the literature, and our model can fit the experiment data with less deviation.