Preference-Based Design in Architecture
# of pages200
ImprintDelft University Press
Architectural design is a complex process as a result of two prominent characteristics
of choice making: 1) multiple designs can fit into one intended purpose, which raises
the question: how to choose the design that fits best, and 2) a multitude of decision
makers have an interest in the design process, which is the problem of group choice
The application of the field of decision theory is aimed at finding tools, methodologies
and software to help people, or groups of people, make better choices.
The scientific foundation of selection (choice) is preference measurement. The correctness
of a decision analysis methodology is determined by the correctness of the scales used for measuring preference. All classical models of the theory of measurement generate scales to which the operations of addition and multiplication are not applicable.
A new methodology called Preference Function Modeling (PFM) offers a correct model
for the measurement of preference and for the selection of the most preferred alternative.
In its current form however, PFM is an evaluation methodology, helping decision makers to choose the most preferred design alternative from a set of already existing alternatives. In the domain of architecture a design methodology is needed, where the design alternatives are not known a priori.
The Preference-Based Design procedure proposed in this thesis offers a design methodology in which the feasibility of considered alternatives is established using the
concept from the Open Design Linear Programming (LP) technique of defining an alternative as a combination of decision variable values within negotiable constraints. The PFM algorithm is used to rank the feasible design alternatives on preference.
The relevance of this thesis is two-fold:
1. The quality of decisions is improved, since only feasible designs are taken into
2. The acceptance of decisions is improved, since all decision makers can see that their
interests are genuinely taken into account.