基于遗传算法的一体化通风网络优化算法

Integrative optimization algorithm of min ventilation networks based on genetic algorithm

概括混合型一体化通风网络优化的模型,分析目前混合型通风网络优化的4种求解方法优缺点。针对混合型通风网络优化的要求,提出混合型通风网络风量分配和风流调控一体化的优化思路。在通风网络理论和图论的基础上,引入遗传算法随机产生2个动态网络的邻接矩阵和余树弦风量值,使用附有条件的最小支撑树算法产生2个最小支撑树,进而求得相应的回路矩阵。通过余树弦风量值和回路矩阵等分别计算通风网络风量分配值和风阻调节值,基于通风总功率和约束条件构建广义最小化目标函数,依此对分风和调风方案进行评价,使用遗传算法中进化算子对分风和调风方案实施进化操作,最终得到满意解。研究结果表明:该算法是严格数学意义上全局优化算法,解决调风地点约束的通风网络优化问题,利用网络结点流量平衡的等式约束条件,减少最优化模型中变量数目,提高算法效率。

The mixing optimization model of min ventilation networks was summed up,the four way advantage and disadvantages of sloving mixing ventilation networks optimization problem were analysed.Facing the demand of mixing ventilation networks,integrative optimization way optimizing min ventilation networks was put forward.The two adjacency matrix of dynamic networks were initialized on random according to genetic algorithm thought based on min ventilation networks theory and graph theory.The two minimum spanning tree of the dynamic network was searched by the way of minimum spanning tree algorithm confined in conditions,two independence circuit matrix were calculated,and then the ventilation volume of networks branchs and resistance adjusting values of remaining tree branchs were calculated by the ventilation volume of remaining tree branchs and circuit matrix.The generalized objective function is set up in the sum ventilation power and restriction conditions,The ventilation volume distributing and adjusting schemes were judged by the generalized objective function,and schemes codes were evolved by genetic operator.Finally,the satisfaction ventilation volume distribution and adjustment schemes was attained by iterative.The algorithm is a globe optimization in strict mathematics define.The algorithm deals with adjusting location restriction very well.The results show that the algorithm reduces variable number in mode,and is good at efficiency,through making use of the node int-flow and out-flow balance restriction.