Multi-objective building design optimization under operational uncertainties using the NSGA II algorithm

Shobhit Chaturvedi, Elangovan Rajasekar, Sukumar Natarajan

Research output: Contribution to journalArticlepeer-review

30 Citations (SciVal)

Abstract

Operational uncertainties play a critical role in determining potential pathways to reduce the building energy footprint in the Global South. This paper presents the application of a non-dominated sorting genetic (NSGA II) algorithm for multi-objective building design optimization under operational uncertainties. A residential building situated in a mid-latitude steppe and desert region (Koppen climate classification: BSh) in the Global South has been selected for our investigation. The annual building energy consumption and the total number of cooling setpoint unmet hours (h) were assessed over 13,122 different energy efficiency measures. Six Pareto optimal solutions were identified by the NSGA II algorithm. Robustness of Pareto solutions was evaluated by comparing their performance sensitivity over 162 uncertain operational scenarios. The final selection for the most optimal energy efficiency measure was achieved by formulating a robust multi-criteria decision function by incorporating performance, user preference, and reliability criteria. Results from this robust approach were compared with those obtained using a deterministic approach. The most optimal energy efficiency measure resulted in 9.24% lower annual energy consumption and a 45% lower number of cooling setpoint unmet h as compared to the base case.

Original languageEnglish
Article number88
JournalBuildings
Volume10
Issue number5
DOIs
Publication statusPublished - 1 May 2020

Keywords

  • Multi-criteria decision making
  • Multi-objective building design optimization
  • Robustness assessment

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Architecture
  • Building and Construction

Fingerprint

Dive into the research topics of 'Multi-objective building design optimization under operational uncertainties using the NSGA II algorithm'. Together they form a unique fingerprint.

Cite this