Deadline for individual submission

The deadline for the individual submission is approaching, so it is time to focus on this and get it done. The deadline is Monday 27th September, 2pm.

There are some key things to remember, starting with the most important:

• Make sure that you iterate!!! The complexity of the form is less important. The most important thing is that you have iterations showing differing levels of performance. Make sure that you extract the data and plot the graphs as we did during the tutorials. We discussed three different types of graphs - 1) parameter sweeps, 2) pareto graphs, and 3) spider diagrams. Use these techniques to understand whether you are improving things, or making things worse. In order to iterate through large numbers of variants, use the animation tools (e.g. the stair CHOP) to develop and evaluate variants automatically.
• When you are considering you evaluation criteria, you need to choose at least two criteria. Try to choose criteria that conflict - for example maximise day VF and minimise sun VF. Then you need to decide what surfaces to analyze - just analyzing everything may not make sense. You will also need to consider carefully how to calculate some global values that represent the overall performance of each variant - do you take the average? or the area weighted average? or just the total? or do you calculate the percentage above a certain threshold?For example, for sun VF you may want to calculate the number of windows with a sun VF of more than 20%.
• Start with a very simple building form!!! I would suggest something like a slab, or a group of slabs, or a tower. Make sure that everything you generate is the right plot ratio. The proportions of these forms should be building like - e.g. don't make apartment blocks that are 30 meters deep. Start by varying these simple geometries and running simulations for these variants. Find the simple form with the best performance. Try to choose parameters that result in forms are still feasible - there is no point in simulating something that you already know will not work anyway. So if you have a slab apartment block, then there is no point in varying the depth much, since you already know that it is very constrained by daylight and ventilation issues.
• Then you can start to make the form more complex to try and improve performance even more. This usually means adding a few parameters to your simple form so that you can generate further variants. Run the simulations again to check if you were right - did performance get better or worse? If it got worse, then try something else but make sure you keep the bad results in your graphs. When making the form more complex, don't over do it - keep it manageable. And remember - we are only interested in the overall massing, and some general ideas of internal layout (e.g. lift cores, circulation) - so don't worry about the handrails...
• Some of you are exploring some more complex form generating techniques, like metaballs and L-systems. But even with these techniques, start simple. Use these techniques to generate a range of building forms, from very simple to more complex, and then simulate to test so that you can understand how performance changes with complexity. So, an L-system can make a straight tower, and a slab block... so start there.
• In order to come up with an idea for your simple form, you may decide to run some initial simulations as a way of understanding the site better. This may involve inserting some simple geometries into the site - e.g. a big cylinder or a big box  (these are not really building forms, since the plan is much too deep). Some of you have already done this and produced some interesting data. Now it is time to start inserting some buildings.

For the submission, you must 1) do a presentation, and 2) submit an A3 landscape booklet covering your presentation. You may also submit a CD with animations if you wish.