THE ROBUSTNESS OF GENETIC ALGORITHMS IN SOLVING UNCONSTRAINED BUILDING OPTIMIZATION PROBLEMS
THE ROBUSTNESS OF GENETIC ALGORITHMS IN SOLVING UNCONSTRAINED BUILDING OPTIMIZATION PROBLEMS
复制标题
遗传算法解决无约束建筑优化问题的鲁棒性
DOI：
--
复制
发表时间：
2005
期刊：
影响因子：
--
通讯作者：
Ali F. Alajmi
中科院分区：
文献类型：
--
作者：
J. Wright;Ali F. Alajmi
本文研究的鲁棒性遗传算法（GA）的搜索方法在解决无约束的建筑优化问题，当优化所使用的建筑模拟的数量是有限的。GA搜索方法可以被分类为基于概率群体的优化器。搜索的概率性质表明，GA在寻找解决方案时可能缺乏鲁棒性。此外，这是一个共同的看法，因为遗传算法的种群（集）的解决方案，他们需要许多建筑模拟收敛。它的结论是在这里，一个特定的遗传算法是强大的，在寻找解决方案与1.4%的平均差异，在建筑能源使用的最佳解决方案，在所有的试验优化。仅用300个建筑物模拟（在任何一个试验优化中）就实现了这一性能。
This paper investigates the robustness of a Genetic Algorithm (GA) search method in solving an unconstrained building optimization problem, when the number of building simulations used by the optimization is restricted. GA search methods can be classified as being probabilistic populations based optimizers. The probabilistic nature of the search suggests that GA’s may lack robustness in finding solutions. Further, it is a common perception that since GA’s iterate on a population (set) of solutions, they require many building simulations to converge. It is concluded here that a particular GA was robust in finding solutions with 1.4% mean difference in building energy use from that for the best solution found in all trial optimizations. This performance was achieved with only 300 building simulations (in any one trial optimization).