In Google Earth, we can find the fabric of the block is the standard type of old urban

Since we consider about urban fabric, we started from figure-ground.

The paths in figure ground could be deconstructed into “network system” and “path themselves”

Network System

Path Themselves

So, let me introduce the process in a brief way.

Path analysisNetwork analysis

Select the target block

Network generation

Path generation

Combine network and path together to get a new block

Remake as Figure-ground

System features

Hierarchy

Cellular-automata

System bug fixing

Reform in new block Reform a standard path by patterns

Take patterns

Pick standard paths

People purposely make this kind of paths, Block could be cut into more than 2 parts by this hierarchy.

Network Analysis

Main

Inferior

Secondary

We start from the hierarchy analysis to define which level could be make by generative design.

With lowest hierarchy, used to support the system of green ones.

Main element of the path system, mainly used by local people, with high continuity.

I want to deal with two hierarchy only, so I take a part as the study area.

In other word, the site of this design is limited in this area.

By the way, if we consider about network system only, the mapping is just like this.

Network Analysis

Diagram-街道續Diagram-障礙移除

People in the past form paths in a simple thought----the continuity.

In the past, paths is made without urban planning, people just try to link different part, or break obstacles to make the continuity of paths.

Network Generation

In this generative design, we use cellular automata to form a new network system.C.A. in this case is a grid system construct with a lot of cells, each cell acts follow a simple rule. Like “When there are 1 cell around, the cell alive.” or “When there are more than 4 cells around, the cell should die.”

We use this kind of rules, to analogue the behavior of path making of past.

Network Generation

We use the plug-in of “Grasshopper,” “Rabbit”

This battery control the rule of born.Dead cell comes to alive when it has exactly 1 or 2 living neighbors.Dead cell comes to alive when it has exactly 1 or 2 living neighbors.

This is the rule how a cell born in this case.

Network Generation

A living cell remains alive only when surrounded by 1,2 or 3 living neighbors. A living cell remains alive only when surrounded by 1,2 or 3 living neighbors.

This is the rule that a cell could keep alive. If the cell couldn’t keep alive, then die.

Network Generation

This Is the process we make network system.

We start from the network made by cellular automate

And we note the network system we

see in the C.A..

Then fix the system in order to make

sense in real.

Define the hierarchy of the paths inside.

(Into blue and green)

And we smooth the green system to

strengthen continuity.

Network Generation

And we transform the network we got in C.A. into the block we deal with

Network Generation

This part, we are discuss what define the form of path on figure ground.I tried to tell a path in many patterns, each patterns have its own width and length.The 2 basic parametric are width and the length of patterns.

Path Analysis

Since I just analogue computer, I decide analyze a path for each hierarchy only.

At the same time, we will use the same analytic process on blue hierarchy, too.

Path Analysis

This analysis is based on how I realize the path made from.

Path Analysis

If we consider the path is the space between the buildings along itself.

We can use the length of building as a parametric and the width of path as the other parametric .

We will got a lot grids as patterns.

Green System Blue System

length width 1/2 width length width1/2

width length width 1/2 width length width 1/2

widthS 1 11.2 2.84 1.42 D 1 12.88 1.21 0.605 S 1 3.28 0.9 0.45 D 1 3.07 0.97 0.485S 2 7.33 3.12 1.56 D 2 4.12 2.49 1.245 S 2 12.27 1.51 0.755 D 2 4.02 3.82 1.91S 3 3.58 3.08 1.54 D 3 4.39 1.68 0.84 S 3 3.81 1.64 0.82 D 3 5.12 1.6 0.8S 4 21.77 3.45 1.725 D 4 4.92 1.44 0.72 S 4 12.79 1.49 0.745 D 4 5.65 1.42 0.71S 5 10.71 2.78 1.39 D 5 3.67 2.72 1.36 S 5 7.4 1.48 0.74 D 5 10.3 3.23 1.615S 6 4.55 3.09 1.545 D 6 12.67 1.83 0.915 S 6 12.13 1.63 0.815 D 6 5.54 1.43 0.715S 7 10.82 4.43 2.215 D 7 12.42 2.81 1.405 S 7 27.57 2.42 1.21 D 7 9.33 3.56 1.78S 8 6.62 3.91 1.955 D 8 8.05 1.92 0.96 S 8 17.18 2.91 1.455 D 8 15.32 1.48 0.74S 9 3.44 3.97 1.985 D 9 2.26 1.64 0.82 S 9 3.73 4.14 2.07 D 9 5.36 2.36 1.18S 10 15.79 4.2 2.1 D 10 5.19 1.87 0.935 S 10 4.35 3.33 1.665 D 10 10.97 1.99 0.995S 11 15.34 3.29 1.645 D 11 2.25 2.19 1.095 S 11 3.29 2.72 1.36 D 11 5.03 2.9 1.45S 12 19.47 3.78 1.89 D 12 13.11 3.03 1.515 S 12 1.29 0.39 0.195 D 12 2.46 1.69 0.845S 13 17.23 3.35 1.675 D 13 9.66 2.98 1.49 S 13 0.76 1.47 0.735 D 13 3.05 1.86 0.93H 1 15.42 3.31 1.655 C 1 1.2 2 1 H 1 1.74 1.2 0.6 C 1 5.21 2.9 1.45H 2 17.8 4.27 2.135 C 2 6.06 1.99 0.995 H 2 2.53 1.53 0.765 C 2 2.23 5.41 2.705H 3 4.71 3.06 1.53 C 3 4.07 2.32 1.16 H 3 1.08 1.14 0.57 C 3 8.12 1.4 0.7H 4 8.34 2.94 1.47 C 4 4.53 3.51 1.755 H 4 1.3 1.38 0.69 C 4 6.59 3.37 1.685H 5 4.14 2.69 1.345 C 5 8.71 2.76 1.38 H 5 2.36 1.37 0.685 C 5 22.37 3.44 1.72H 6 10.69 2.73 1.365 C 6 16.27 2.56 1.28 H 6 3.78 1.77 0.885 C 6 4.82 1.4 0.7H 7 10.56 2.42 1.21 C 7 3 2.71 1.355 H 7 4.61 1.16 0.58 C 7 3.25 6.1 3.05H 8 5.15 3.52 1.76 C 8 1.43 1.31 0.655 H 8 4.77 1.19 0.595 C 8 3.57 3.98 1.99H 9 8.29 2.62 1.31 C 9 3.54 4.15 2.075 H 9 0.76 1.71 0.855 C 9 7.16 3.57 1.785H 10 12.49 1.71 0.855 C 10 4.17 3.07 1.535 H 10 5.9 0.39 0.195 C 10 5.46 2.57 1.285H 11 4.45 3.18 1.59 C 11 12.96 3.21 1.605 H 11 6.29 0.97 0.485 C 11 10.91 2.57 1.285H 12 7.37 4.2 2.1 C 12 0.45 2.73 1.365 H 12 3 1.15 0.575 C 12 5.03 2.32 1.16H 13 26.62 3.97 1.985 C 13 4.28 3.54 1.77 H 13 1.99 3.28 1.64 C 13 No Data

joker 1 4.19 2.61 1.305 joker 1 No Data

joker 2 5.09 4.53 2.265 joker 2 No Data

S = 　 Space D= 　 Diamond H= Heart C= 　 Club

♠ ♦ ♥ ♣

You can see we got a lot of data.

Since I use card as my random picker, my data of each hierarchy should controlled under 54 patterns.

Path Analysis

To form a new path, we put a lot patterns together.

Path Generation

The　last　process!

And now, we transform the paths into the network, to put them together.

This process really need the help of computer…

Finally, we got a new path… But there is still something different…

Compare　with　old　one…

I think it’s not difficult to find the new network is

too smooth, and the width of paths have less

variation.

Finally, we got a new path… But there is still something different…

Compare　with　old　one…

If only we compare with the path we analyzed, the outcome

is ok.

Because the path we choose is obviously working in the whole.

So…　How　can　we　make　this　process　better?　

Just like debug, each time we run a process, we can know what might fix this process.

Maybe there are more hierarchies. These two paths in my analysis are at the same hierarchy, but the form of them are really different.

To find the differentiation, we should know more reasons what make the form

of paths.

And　we　can　still　find　the　path　is　really　not　smooth.

We also need more obvious about the reasons form the network.

some place smooth, so place angle

Full of angles Too smooth

The real oneCellular automata type Smoothen type

To make a real block by generative design, we need more data (parametric) to get closer to real.

Path

Network

And more?

Thanks for your attention.

Always　end　with…

THE　　　　　　　END