中文摘要：

着眼于AEA算法本身的不足点，通过更合理地给定其每一代的行走步长，提出一个改进的AEA算法（IAEA）来提高AEA算法的寻优性能。鉴于行走步长对于进化的不同阶段全局搜索和局部搜索之间的关系影响，今提出IAEA的步长应随着实际进化的不同阶段而合理地变化，以使得算法能跳出局部最优，避免早熟现象的发生。IAEA算法在7个典型测试函数上进行了测试。测试结果表明，与基本AEA算法、PSO和DE算法相比，IAEA的寻优性能有了很大的提高，不仅获得的解的质量更好，而且算法的稳定性都得到了提高。最后将IAEA算法用于重油热解模型参数估计的仿真研究中，通过验证，得到了更有利的结果，说明文中提出的算法是有效的。

英文摘要：

Focusing on the deficiencies of Alopex-based evolutionary algorithm （AEA）, an improved AEA algorithm （IAEA） was proposed, in which a more rational moving step length to each generation was given. As the moving step length plays a key role on the relationship between global search and local search in different phases of the evolutionary process, this paper proposes that the IAEA＇s step length should be changed reasonably with the actual evolutionary process, so that the algorithm can jump out of local optimal value and avoid premature phenomena. Then the IAEA algorithm was tested in seven benchmark functions, and was compared with PSO, DE and AEA. The promising results show that the IAEA algorithm not only obtains the solutions with much better quality, but also with higher stability. Finally, the IAEA algorithm was applied to a parameter estimation of heavy oil thermal cracking three lumps model, and good results were achieved. This shows that the proposed IAEA algorithm is very effective.