Delaying the analysis of the building performance to the final stages of the design process can often cause the actual results to differ from the expected ones, with the issue of significant redesign costs. The aim of this work is to improve the designer’s control on thermal and geometric variables of the building in the first stages of the design process, to obtain an optimized building from different points of view. An innovative computational performance-driven design optimization workflow is proposed. The integration of a parametric algorithmic modelling tool in a genetic algorithm powered performance optimization procedure allows to minimize energy need and construction cost of the building, and to optimize the natural illumination levels, affecting several geometric variables. This approach is tested on the early design stages for the new construction of an educational building located in the centre of Italy (L’Aquila). The results are post-processed in order to provide an innovative graphical visualization of the outputs to improve the decision-making process. Useful insights on the influence of different design choices on three key performance indicators can be drawn for the case study building, by comparison of the Pareto solutions: optimization results highlight the strongest influence of window-to-wall ratio, among the other variables, on energy need and useful daylight illuminance.

将建筑性能分析推迟到设计过程的最后阶段通常会导致实际结果与预期结果不同，并带来大量的重新设计成本。这项工作的目的是在设计过程的最初阶段改善设计人员对建筑物的热和几何变量的控制，以便从不同的角度获得优化的建筑物。提出了一种创新的计算性能驱动的设计优化工作流程。将参数算法建模工具集成到遗传算法驱动的性能优化过程中，可以使建筑物的能源需求和建筑成本降至最低，并优化自然照明水平，从而影响多个几何变量。这种方法已在位于意大利中心（L'Aquila）的教育建筑的新建工程的早期设计阶段进行了测试。结果经过后处理，以提供输出的创新图形可视化效果，以改善决策过程。通过比较帕累托解决方案，可以得出有关不同设计选择对三个关键性能指标的影响的有用见解，这些案例可通过比较帕累托解决方案得出：优化结果突出了窗墙比对其他方面的最大影响。能源需求和有用的日光照度。