DEVELOPING A COMPUTER ASSISTED MULTIDISCIPLINARY DECISION MAKING PLATFORM

1
DEVELOPING A COMPUTER ASSISTED MULTI-DISCIPLINARY
DECISION MAKING PLATFORM

• 2008 Structures Congress Crossing Borders
• Vancouver, Canada
• Friday, April 25, 2008
• Jason E. Charalambides

2
ISSUES ON COLLECTIVE DECISION MAKING

• In the field of construction, a number of
  individuals pursue the option of making
  decisions each from their own point of view.
• Reasons for disagreement are not the ultimate
  success of a project but the priorities and the
  method to be followed. Of course we all want to
  make a project succeed but how do we do that?
• By being affixed to our own positions, based on
  our own body of knowledge, it is anticipated that
  we shall not agree with whoever is on the other
  side of the table.
• Dynamics develop among groups and individuals.
  Architect disagrees with code inspector, Engineer
  has issues with Construction manager etc.

J. Charalambides
3
ISSUES ON COLLECTIVE DECISION MAKING

• This leads to a reduction of the collective
  capacity over the capacity of the individual,
  bringing to mind the slogan Decisions should be
  made by those who specialize on a subject and not
  by the entirety of the citizens.
• In the end we see decisions made on the basis of
  articulation of arguments and personalities of
  players.

J. Charalambides
4
ADVANCES IN TECHNOLOGY

• Computer Aided Design of the mid 80s when it
  started becoming a common practice, was nothing
  more than a basic interpretation of drafting
  within an electronic environment.
• We were comfortable with the idea of 2D drafting
  and as it is easy to reprint a sheet of drawings
  with any changes or corrections, we found
  adequate gain to justify investing in computer
  technology.

J. Charalambides
5
DO WE USE TECHNOLOGY?

• The capacity of Information Technology was
  cornered to a very advanced area of vector
  drawings, while the basic notion of simple
  statistical relationships that can yield
  suggestions or even final decisions were not
  addressed.
• Instead, we hold meetings!
• Yet, failure of any project will likely occur
  because of system failure, that of human
  interactions and dynamics, not by extraneous
  factors.
• Columns do not buckle because of excessive load.
  They buckle because they were not designed to
  assume the excessive load.

J. Charalambides
6
DECISION MAKING

• The case of a column failing is very easy to
  address and it would be difficult to miss.
  Similar would be a case of designing a building
  according to code.
• Incorporating a series of such parameters within
  the design process is what CAD should have
  already been since its inception. It should be an
  ever-present factor that aids toward the design.
• Nevertheless, we should question where we are
  standing with respect to this early step of
  implementing decision making within the framework
  of design.
• That would be the gradual relief of human
  intervention, or the permission of voluntary but
  not mandatory interaction between computer and
  designer.

J. Charalambides
7
APPLICATION LEVEL I

• At a primary level, the use of algorithms may be
  applied for an automated/interactive and
  parametric method of generation of structural
  forms. Process is linear and parameters are
  clear.

J. Charalambides
8
APPLICATION LEVEL II

• Scenario
• Design of a Tensegrity structure Choice of
  materials
• There is a 70 chance steel will cost twice as
  much as aluminum, and 30 chance it will be three
  times as expensive
• For steel, labor costs are approximately equal to
  material cost for aluminum, labor costs 2.5
  times as much as material cost.
• For aluminum systems there is a 15 chance that
  preassembly will be overly pre-stressed. If this
  occurs there will be a 60 consideration of total
  replacement of the compressive elements.
• With 25 chance of connections being available
  only in steel, insulation of steel from aluminum
  may fail at 20 of the cases. That will lead by
  90 to corrosion and will force a total
  replacement. Cost of total replacement will be
  based on a normal distribution with a mean of
  140 of original cost and s? 20.
• For steel there is a 3 likelihood that some
  elements will fail prematurely, with any
  components costing 80 of the original cost.

Image Tensegrity model built by Dr. Katherine
Liapi at the University of Texas at Austin, Dept.
of Civil, Architectural and Environmental
Engineering.
J. Charalambides
9
APPLICATION LEVEL II

• At the first level the amount of time needed is
  destined for the generation of an actual form.
  However, the parameters are easy to establish,
  and with some visualization aid, fine tuning of
  forms can take place in negligible time.
• When risk analysis and decision making should be
  applied is when all the parties involved may
  begin to question the process.
• A simple example for deciding the material that
  will be used for the previous structure was
  developed, and a Monte Carlo simulation was
  modeled to produce a final result. Data were
  given and they provided no grounds for
  argumentation. Information Technology integrated
  to CAD minimized the role of a meeting to a
  mere opportunity for information on the process.

J. Charalambides
10
APPLICATION LEVEL III

• Assessing the value of quality
• No value is to be considered subjective. In an
  interactive manner, values may be applied toward
  factors that will determine quality levels.
• The parameter of visual complexity may be broken
  down to the six elements that define it
• Color,
• Form,
• Light,
• Shadow,
• Texture and
• Contrast.
• All of the above can be given values, and the
  subjectivity of qualitative values will be
  quantified.

Comparison of interest and liking versus
complexity. Rapoport, Amos Human Aspects of
Urban Form, 1977.
J. Charalambides
11
CAN THE COMPUTER PERFORM THAT LEVEL OF ANALYSIS?

• Is all that information available to computers?
• In what language do we see the codes written?
• Should we all invest in BIM?
• It depends on the clientele a company is
  addressing.
• We will all be using BIM at the stage when we
  shall not engage into the details.
• Do we all remember how we used to connect to the
  internet? Or maybe how we used to use
  Wordperfect as a word processor?

J. Charalambides
12
CAN THE COMPUTER PERFORM THAT LEVEL OF ANALYSIS?

• A computer becomes an entity that analyzes data
  and yields results. Once data is available, work
  can be automated, minimizing decision making and
  implementation process time.
• At that stage, a computer becomes an autonomous
  decision making agent.

J. Charalambides