一种改进编码策略在演化函数优化中的应用

Implementation of Improving Encoding Strategy in Evolutionary Function Optimization

在用演化算法求解函数优化问题时，以往常用的二进制整体编码存在着诸如位串过长、操作重复、执行效率不高等一些弊端。特别是对高维问题，这些弊端就表现得更为明显。针对这些弊端，文章给出了一种改进编码－独立编码策略。通过对大量测试函数的实验，证明这种编码确实能改进上述缺点。另外，通过独立编码策略，还提出了一种基于问题解空间分解的并行模型。实验证明，这种并行模型能够得到较高的线性加速比。

In the field of evolutionary function optimization, there are some shortcomings mostly caused by using the binary enttirty encoding,they are:too long encoding, repeated operation, slow efficienncy and so on. Specially to high-dimensional functions, these shortcomings become more obvious.In order to overcome them, we offer an improving encoding-independent encoding. A lot of numerical experimental results show that this encoding can really improve those shortcomings. Otherwize, through independent encoding, we offer a new parallel model according to the resolvability of independent variable room. Experiments prove that this model can get higher linear speedup.