You can simply start using the Wallacei Analytics tab by clicking on ‘Draw’ at the bottom left. The results of the optimization will show.  

You can simply start using the Wallacei Settings tab by clicking on ‘Draw Parallel Coordinate Plot’ at the top left. The results of the optimization will show.  

When you open it is empty. To see, select a solution you want to analyse. You can start with the last produced solution. In this case 24,24, click on Draw.  

You will see this. Alternatively, you can also show where in the analytics space the solution is located. Click on ‘Select’.  

It looks as follows. In the next steps we will go into all the 5 diagrams to see what they mean and how you can use them.  

Sometimes referred to as “star coordinate method”, the Diamond Chart analyses the fitness values of single solution, and not the whole population. It shows how a single solution performs in terms of its individual values and fitness objectives, helping the user to understand those relationships in more detail for chosen solutions. 

The Standard Deviation Graph’s main purpose is to show how much the solutions change over iterations for each objective. Each curve represents the range of solutions in a single generation/iteration. The narrower the curves, the more stable and consistent the solutions are. The wider the curves, the more variation there is. In this example, you can see that the WWR becomes the most consistent and stable over time, while the unique panels’ objective is wider meaning it’s more varied and it’s having a harder time finding a stable solution. This graph helps track how the optimization process refines solutions over time. 

The Fitness Value Graph  tracks the performance of the solutions over the iterations for each goal. Each line represents the fitness values of each solution across the iterations, basically checking how well they are doing to reach that objective  over time. For the WWR and the cost, you can see that the solutions are more scattered in the beginning, but they eventually improve more towards our goal. However, the values keep fluctuating for the unique panels  objective, clearly showing that its more difficult to achieve that objective. This graph helps visualize how each objective’s solutions evolve and whether they converge toward optimal results or continue exploring diverse possibilities.  

The SD Value Trendline tracks how the standard deviation of each objective changes across the different iterations, showing whether the solutions in the population are becoming more stable or remain varied. For the WWR objective, the trendline quickly flattens, indicating that the solutions coincide and stabilize early. In the cost’s objective, there is also an initial drop followed by some fluctuations. This shows that it took a little more time for solutions to stabilize. For the unique panels’ objective, the standard deviation keeps fluctuating throughout, showing that variability remains, the algorithm is still trying out different options and hasn’t fully stabilized. This graph provides insight into the optimization process by highlighting when each objective achieves stability or continues evolving. 

The Mean Value Trendline tracks the average performance of solutions across iterations for each objective. In the WWR objective, the initial values are scattered but quickly stabilizes, showing rapid progress toward optimization. The cost objective starts with high values and gradually decreases, stabilizing as more cost-efficient solutions are found. For the unique panels’ objective, the values decrease steadily, but some fluctuations remain, suggesting the algorithm is still exploring various configurations. This graph helps visualize how each objective’s solutions evolve and whether they are converging toward optimal values or continuing to explore diverse possibilities. 

In the “Parallel Coordinate Plot” you can see solutions as lines, going up and down depending on each objective fitness. Red means the oldest, and Blue means the newest. It helps understand how the solutions relate to each other, and how each fitness values influence others, or not. 

The Parallel Coordinate Plot can be analysed in the Wallacei Selection tab. When you open the Wallacei Selection tab, all the graphs will be empty. You can start with looking into the parallel coordinate plot. Draw the parallel coordinate plot by clicking on the button in the left top corner.  

The coordinate lines will appear, which you have seen while the optimization algorithm ran. Now it is time to investigate these results.  

The first Analysis Method is “repeated Fitness Values”, which shows how often, depending on the objective chosen, a certain value occurs. Check the box of ‘Parallel Coordinate Plot (PCP) Settings’ and change the analysis method to ‘Repeated Fitness Values’. To show the most repeated fitness values of an objective you can write the number of the objective you want to investigate and click on “Run PCP Analysis”. 

After you have investigated the most repeated fitness values per objective, you can dive deeper into the analysis by selecting the solutions with the most repeated fitness value to see how the solutions are located on the parallel coordinates plot. Change the analysis method to ‘Solutions with Repeated Fitness Values’. Next write down the objective number and ranking of the solution you want to investigate.  

The selected raking with the solution will show on the PCP, highlighted in black, and you can find the analytics results on the left side. It is possible that the selected ranking of repeated fitness values results in different objective fitness for the other objective, if this is the case you will see multiple lines in the PCP.  

After you click the “Run PCP Analysis”, you can see in the next window to the right all the solutions with the repeated fitness value. From this list you can select solutions and add them to your export list. 

Change the analysis method to ‘Relative difference between fitness ranks and write down the ranking of the solution you want to investigate. 

Change the analysis method to ‘Average fitness values’ and write down the ranking of the solution you want to investigate. 

You can also investigate generations on the Parallel Coordinate Plot. Check the box ‘Generations’. In the solutions list all the generations will show, here you can select one generation and add the generation to the export list.  

In the control panel you can also select one generation and show the generation on the PCP.  

The Objective Space (OS) and Pareto Front (PF) components are recommended to be used together. While Objective Space can be used on its own, Pareto Front component needs the Objective Space to work. The Objective Space component displays the objective space of all generations. The Pareto Front component calculates the non-dominance value and draws the Pareto Front for any given generation. This is incredibly helpful when analysing the solution space, to find the “most optimal” solution.  

In the first tab of Wallacei X, the Wallacei Settings tab, you can visualize the pareto front in the objective space.  

Let’s assume you opened a saved Wallacei optimization component that includes the optimization results. In the objective space window (bottom right) click on the “Draw Entire Population” button.  

You can see all the generated options in 3D space, on axis depending on each objective fitness. The yellow boxes are highlighting the Pareto Front solutions, you can see them as essentially a range of the fittest solutions. In the objective space you can see the convergence of the solutions towards the optimum.  

You can also show the pareto front in the Wallacei Selection tab on the Parallel Coordinate Plot and select solutions for export.  

When you open the Wallacei Selection tab, all the graphs will be empty. Draw the parallel coordinate plot by clicking on the button in the left top corner to start the analysis.  

Check the box of ‘Pareto Front Solutions’ to start. After you have checked the box, in the second tab the solutions of the Pareto Front will appear for you to select them for export.  

You can show the Pareto Fron on the PCP by clicking on ‘Show Pareto Front on PCP’.  

You can now see Pareto Front solutions highlighted in black on the PCP. You can notice, that in our case, those are just the members of the last generation.  

You can cluster a generation by checking the box of ‘Unsupervised Machine Learning’ and select the ‘Kmeans’ clustering method. Input the generation you want to cluster and the number of clusters you want to generate. Then click on ‘Run’.  

After clicking on run the results of the clustering will show in the Wallacei Selection tab. In the right bottom you will visually see the clustering in the objective space plot.  

Each cluster is represented by its own colour. Investigate the clustering in the Objective Space, it is important to check if the clusters are made logically. If you click on ‘Run’ again in the Control Pane, the clustering algorithm will create new clusters.   

The number of clusters you want to create highly influences the outcome of the clustering. You will have to look critically to the results to see if the number of clusters is relevant for the problem at hand. For example, too many clusters for too little solutions will simply cluster duplicate solutions together. While too little clusters could result in similar solutions being grouped in different clusters and adjective solutions being grouped in the same cluster

In the second window from the left, you will find all the solutions of the generation grouped per cluster.  

You see that certain solutions are highlighted bold. These bold solutions are the centre of the cluster. This solution represents the cluster.  

If you want to compare clusters, it is worth investigating the centres of each cluster. You can add those centres to the export list.  

You can also visualize the results of the clustering on the Parallel Coordinates Plot (PCP). In the Control Panel click on ‘Show on PCP’.  

The results of the clustering will appear on the Parallel Coordinate Plot in the top right window. Each cluster is represented by its own colour and the centre of the cluster is highlighted with a bold dotted line.  

First make sure that the components holding geometry you want to export are plugged into the WallaceiX component. Minimize the Wallacei X tab and check if you have geometries and data connected to the ‘Phenotype’ input and the components are enabled.  

The next step is to export the solutions in the Wallacei X interface. In the WallaceiX interface, go to the Selection Tab, click on the “Export” button. 

The export will start loading now. A pop-up will show stating the number of exports. Click on ‘OK’ to start the export.  

The time the export takes depends on how many solutions you selected and the geometry that you want to export. In the export list you can see the progress of the export.

Once the export is finished a new pop-up will show, stating the export is finished. You are now done in Wallacei X. Click on ‘OK’ and close Wallacei X.  

Now, go back to the grasshopper canvas interface. You can plug in a panel to the “Phenotypes” output to see what’s inside. As you can see, the data is stored as a “Wallacei Phenotype”

To save the exported phenotypes, create a ‘data’ component and internalize the data. Next save your Grasshopper file to save all the progress.  

You can now safely close the Grasshopper script before moving on the next steps. The visualization of phenotypes is explained in the next tutorial.