The last step of the workflow comes once abase case scenario has been modeled, and the daylight and thermal simulationshave been tested. This task is performed by the plugin Octopus, which is amulti-objective optimization tool. As Robert Vierlingerstates, “Octopus is a tool which introduces multiple fitness values to theoptimization.
Providing a set of possible optimum solutions that preferably reachform one extreme trade-off to the other are probed through the best trade-offsbetween the objectives” (Vierlinger 2014).One of the advantages of using octopus incomparison with other similar optimization tools in Grasshopper is the abilityto define multiple objectives that can be evaluated simultaneously whichprovides optimal results, presenting the best mode of objectives. This approach ofoptimization enables the examination ofcorrelations between different objectives and provide a more comprehensivearrangement of outcomes compared to single objective optimization analysis(Deb, 2001).
The optimization studies included on runningthree generations (sets) with a population of 30 elements each to determine thebest correlations between sUDI and EU. The result of multiple iterationsperformed by the optimization analysis are scattered through athree-dimensional graphic that presents the results obtained by the analysis. Afterrunning several simulations, the graphic gets populated with the besttrade-offs between the different objectives, conforming a well-definedarrangement of the boundaries in which all possible best trade-offs couldoccur. Ordering the results inthe graph leads to the Pareto Front (Fig.4).
A point to be mentioned is that optimizationstudies are not comprehensive and that means necessarily the most optimalanswers has not been achieved. The design variables (parameters) that isintended to simulate were variables relating to the windows: window to wallratio, number of windows, window height, sill height. octopus through repeatingthe various arrangements of the design variables, offersthe most efficient options.In the graphical representation of the optimization,one can find patterns within the multiplicity of results, because of that,further speculations about the correlations between the different objectivescan be elaborated.
For example, a direct correlation between daylighting andenergy consumption was found, Fig.4 describes how any increments in UDI areaccompanied by a linear reduction of EU. This directly occurs because lessenergy is consumed for electric lighting with any increment in daylighting.