遗传规划算法在化合物设计、筛选研究中的应用

采用计算机科学中新兴的遗传规划算法思想，结合化学物质的本质特点，运用进化操作来实现化合物的合成设计和筛选。文中针对算法运用讨论了函数集、终止集问题，通过计算元素组成的字符串的化合价的结果来确定适应度函数，既符合化学学科的本质规律，又满足了算法的要求。通过复制、交换和突变操作，经过多代次的进化终止，取得了满意的结果。从实验可以看到，经过遗传规划算法的操作，得到了很多合理的化合物设计和筛选结果。所以说，该方法的应用有效地促进了化合物实验合成和筛选。文章还针对其实用性，从化学本质出发，提出了建议和研究方向。可以说本文是遗传规划在化学化合物合成筛选中运用的成功探索，同时也为进一步研究打下了坚实的基础。

Application of genetic programming in study of design and sieve of the compounds

Using Genetic Programming to solve compound synthetic design and sieve is a feasible method. In this paper, the application is discussed from the function set and terminal set symbolic selection based on the character of the compound. The fitness function is decided and computed through the valence of the string. Three primary operators are selected according to the practice. The satisfying result can be obtained by this efficient operation using GP. It is obvious that this method can be succeeded with the discussion further.