双参数交叉影响的连续域蚁群算法设计*

针对土木工程领域中的复杂参数反分析问题，基于常规蚁群算法进行了数学模型的构建、算法结构分析，并采用残留信息素数量限制、信息素的持久性系数自适应控制和全局更新规则对算法进行了加强设计，提出了双参数交叉影响的连续域组合优化蚁群算法；同时通过选取五个比较敏感的控制因子：蚁群数量、算法收敛标准、最小信息素持久性系数、循环次数和信息素强度常量进行了数字仿真实验，提出了算法的优化组合参数。通过实例对这种参数识别方法进行了验证，理论结果与实测数据吻合较好，表明了算法的有效性，实现了蚁群算法在土木工程连续域问题方面的应用，丰富了蚁群算法的内涵。

Ant colony algorithm design of intercrossing affected double variable for continuous optimization problem

Based on general ant colony algorithm (ACA) to resolving complex calculation parameters problem in civil engineering, this paper developed a new ACA of intercrossing affected double variable for continuous optimization problem. At the same time, conformed mathematical model and arithmetic structure of the algorithm considering characters of practical problem, drew three affected factors including restriction of remained pheromone number, auto-adapted control of enduring coefficient of pheromone, the whole renew rule into the design of the algorithm. Brought out optimizing combination parameters of the algorithm by numerical simulation experiments of ant numbers, arithmetic convergence standard, minimal permanence coefficient of pheromone, circulation times and intensity constant of pheromone. At last, implemented the modified ACA to reverse calculation parameters of subgrade. The result of simple example shows that the modified ACA can solve the reversion problem efficiently. It realizes the application of ACA in civil engineering for continuous optimization problem.