Civil Engineering Reference
In-Depth Information
needed to understand whether the building occupants and building owners are
satisfied with the overall retrofit result.
This chapter aims at providing an overview of recent research and development
in the third phase, that is, identification of retrofit actions by paying special
attention to the methodologies using multi-objective optimization and genetic
algorithms.
2 Building Retrofit Methodologies and Strategies
Nowadays, a great number of innovative technologies and energy efficiency mea-
sures for building retrofit exist. The main issue is to identify those that will prove to
be the more effective and reliable in the long term. When choosing among a variety
of proposed measures, the decision maker (DM) has to reconcile environmental,
energy-related, financial, legal regulation, and social factors to reach the best
possible compromise to satisfy the final occupant needs. In practice, seeking such a
solution is mainly attempted via two main approaches (Diakaki et al.
2008
).
In the first approach, an energy analysis of the building is carried out and
several alternative scenarios predefined by a building expert are developed and
evaluated mainly through simulation (Krarti
2000
). Although many sophisticated
energy simulation programs (e.g., TRNSYS, Energy Plus) are valuable to study the
impact of alternative scenarios on building performance, the iterative trial-and-
error process of searching for a better retrofit action is time consuming and
ineffective due to inherent difficulty in exploring a large decision space.
The second approach, which is the focus of this chapter, includes decision-aid
techniques that are usually combined with simulation to assist reaching a final
decision among a set of alternative actions. In this chapter, a conceptual distinction
is made between multi-criteria and multi-objective models, according to the sci-
entific literature. In a multi-criteria model, the finite set of alternatives (e.g., three
different types of windows) is explicitly known a priori, in general predefined by
the building expert, to be evaluated according to multiple (quantitative and/or
qualitative) criteria that may be expressed in different types of scales. In multi-
objective optimization (mathematical programming) models, the set of feasible
solutions (e.g., the thickness of the wall) is implicitly defined by the decision
variables and the constraints, and the evaluation aspects of the merit of those
solutions are operationalized through objective functions to be optimized.
Jaggs and Palmar (
2000
), Flourentzou and Roulet (
2002
), and Rey (
2004
)
proposed MC-based approaches for the evaluation of retrofitting scenarios. Ka-
klauskas et al. (
2005
) developed a multivariate design method and multi-criteria
analysis for building retrofit, determining the significance, priorities, and utility
degree of building retrofit alternatives and selecting the most recommended var-
iant. Juan et al. (
2009
) developed a genetic algorithm-based decision support
system for housing condition assessment, which suggests optimal retrofit actions
considering the trade-off between cost and quality.
