组合蚁群算法及其化工应用

针对连续空间优化问题,提出基于新型蚁群算法和模式搜索策略的组合蚁群优化算法。该算法将解空间的每维变量都划分成若干子域,根据每维变量各个子域中信息量占每维变量总信息量的比例来决定蚂蚁在各个子域间的转移,并在各子域中引入遗传操作实现蚂蚁品质的提升。同时,当最优解经过若干代没有改进时,对所有蚂蚁通过模式搜索策略加快收敛进程。以非线性连续优化问题为例进行仿真,结果表明:该方法比遗传算法具有更好的性能。最后,将该算法应用于反应动力学模型参数估计,取得良好的效果。

Combined Ant Colony Algorithm and Its Application in Chemical Engineering

A combined ant colony algorithm,which is based on the novel ant colony algorithm and the pattern search,was proposed to solve the continuous space optimization problem.The algorithm divides the space of each variable into several sub-domains and transfers ants between sub-domains of each variable according to the proportion that is the pheromone in the each sub-domain of each variables comparing to the total pheromone of each variables.It adopts the genetic operation in each sub-domain to improve the character of the(algorithm.) Further,the pattern search is used to quicken the convergence course of the(algorithm) when the optimal value hasn't been improved coffer several generations.The simulation result for non-linear continuous optimal problem indicates that the algorithm has better performance than genetic(algorithm.) Finally,the effect of simulation is satisfactory when the algorithm was used for the parameters estimation of reaction dynamics model.