Source: http://cumincad.architexturez.net/system/files/pdf/cb42.content.pdf
describes the work of students
“Artificial life”, is the study of algorithms that try to explain the patterns and structure that we observe in the world, as the emergent properties of parallel behaviours of autonomous processes, which are very
simple. The complexity happens because of the interactions between the processes”
pg 652
“The interaction between the agent or learning system and the environment can be represented as a form of architectural designing, as a set of simple parallel rules from many interacting virtual designers. The trick is to come up with simple but useful representations and rules for the virtual designer to use in interacting with the environment. What is inscribed on the heart of every AL experimenter is ‘simple rules – complex outcomes’. Baroque rulesets are self-defeating, they already specify the majority of the problem solution, and anyway they take too long to program. much better to write small programs that you can chuck away and try again with another terse summing up of the way to behave.”
“W Gray Walter’s (Walter 55) original 1948 robots ‘Elmer & Elsie’ (which he dubbed M.Speculatrix in cod Latin) were equipped with two (thermionic) valves, a motion sensor, and a photosensitive cell. They tended to move towards the light, but away from objects that they bumped into ( and bright lights), they also kept track of their battery charge, and as this dropped the relative strength of the light tropism increased, guiding them back to the battery charger (which had a light on top). He noted the emergence of
“intelligent” behaviour with these simple feedback loops”
So Walter influenced Braitenberg’s vehicles.
pg 653
1994, Dr. Christopher Langton – Swarm. “The intention was to create a general purpose simulation tool for the investigation of concurrent, distributed systems. Swarm … was used in such diverse areas as chemistry, economics, physics, anthropology, ecology and political science”
“The basic unit of the simulation is a swarm, organising a collection of agents in the form of events. Thus the swarm represents an entire model: it
contains the agents as well as the representation of time. In addition swarms can themselves be agents and an agent can also itself be a swarm. This structure allows the building of hierarchical/multi-level models.”
observer agents collect data while simulation is running. Output graphs or files for later analysis.
Swarm was implemented in Objective C
“The Star Logo software (MIT 97) written as a parallel computation machine which provides another powerful experimental device for exploring and
demonstrating the effects of massively parallel populations of interacting agents in biology, physics, geometry, social systems and ecologies.”
pg 654
“The Star Logo software can be used with both ‘turtles’ – autonomous programmable agents who move about, and “patches” which are a grid of squares that make up a cellular automaton on which the turtles move.”
“Waldemar Bandosz (diploma student 96-98)experimented with both turtles and patches to explore emergent movement patterns and land use on the Stratford railway site”
used ant pheremone model to discover emergent streets.
“Pablo Miranda (1998-) also used starlogo to verify and then extend the scope of the spatial analysis described in Bill HIllier’s ‘space is the machine'”
“Research has shown that the geometrical configuration is enough to explain many of the factors influencing human use of space. The idea (as explained in the book) is to chop up the reachable space of a building or urban area into a grid of small squares, which are then treated as nodes in a network or lattice of points. Calculations are made on the number of steps needed to get from each node to all the others, and the cumulative totals are used
to label each node with a value which corresponds to the average number of steps needed to reach all other nodes per node.”
pg 655
“Mike Batty (Batty 94) has reported on the use of starlogo using Diffusion Limited Aggregation – a model of growth first defined by Alan Turing in the 50’s
“Paul Coates worked with A.Khudair (MSc student 1998-1999) on developing the rules for a cellular automata model of the growth of traditional cities in the Yemen.”
pg 656
Coates implemented 3d agents in microstation.
defined “a set of agents with behaviours such as Looker: Constructor:Mover, and feedback loops between agent > model, model > agent, agent > agent. The emergent outcome was the deformation of the NURBS surfaces into
enclosure/ structures.”
“The conceptual framework of this simulation was oriented on the interdisciplinary theory called radical constructivism. This new science is concerned about the relationship between knowledge and reality. One of the main principles is the thesis that human cognitive systems can only compare perceptions with perceptions, so that there is no direct access to the
‘real world’. In this view, humans are not passive lookers on a fixed reality, but active constructors of a world.”
agents modified a rectilinear NURBS environment by moving points on the surface by a calculated vector.
pg 657
“Michel Mesut, Sean Macmillan & Rio Sibbe worked on a 3D CAD model of the Hackney Dalston corridor using agents with behaviours such as builders of
entrances, rooms, utilities, feedback between agents”
pg 658
“Generally speaking the role of the agents was to find suitable places for different types of urban intervention. The agents are started off at chosen or random points and headings in the model and told to walk along the
ground, not through buildings”
“1) As agents move they leave a trail (coloured line) behind them
2) If an agent senses its own trail, then it remembers, and adds up how many it has found
3) After reaching a threshold, the agent deposits a marker (coloured circle)
4) If an agent meets a marker it may (depending on other rules) make an intervention”
the agents are also influenced by datasurfaces which encode information related to “Crime, Fear (subjective!), Density per landuse, Thresholds,
Accessibility” in the heightmap.
pg 659
“With the notion of a scientific approach towards architectural and urban design certain questions have to be raised. What is the main concern of a designer? How can these concerns be achieved in a scientific way?”
“A cognising subject (student) setting up the interplay of these operables and operations and observing the results is making some kind of scientific
experience. So the conscious and intentional making of experiences may be characterised as a kind of experiment. The outcome of these experiments can
be called operational knowledge. This knowledge is always dependent on certain constraints/ preconditions/presuppositions given at each point of
the experiment and also shaped by the culturally disciplined student him/herself.”