Updated: Sep 7
In ICASP13 we have presented the "Holistic Design Platform (HDP) for sustainable and resilient building design": it represents a novel data-driven human-centric digital twin tailored to the smart city.
Over 50% of the word population is currently living in cities, and over the next 30 years it is expected that 70% of the world population will live in cities. There is the need to pursue safe, liveable and sustainable cities. The sustainable development of the urban communities require a holistic approach involving notions of sustainability and resilience. This holistic approach needs a Holistic Design Platform (HDP) under uncertainty able to satisfy the main pillars of sustainability and resilience, and adopting their main tools: energy simulation, lifecycle analysis, decision making under uncertainty, in addition of course to the safety toward extreme events.
The HDP is a societal Cyber-Physical System based on Performance Based Engineering (PBE) approaches, and Bayesian Networks (BN). Its main feature is the comprehensive consideration of the uncertainties within an holistic view during the lifecycle. It is composed of several submodules, including Uncertainty Quantification (UQ) and Machine Learning (ML), Risk Analysis, Decision Making under Uncertainty, Optimal control.
The modules of UQ and ML include:
1) Evaluations of distributions, including their tails, from small sample of data, using the Kernel Density Maximum Entropy (Alibrandi & Mosalam 2017);
2) Data-driven damage assessment;
3) Vision-Based damage assessment through Deep Learning (DL) and Transfer Learning (TL);
The module of Risk Analysis includes:
1) Tail Probability Equivalent Linearization Method;
The module of decision making under uncertainty includes:
1) Expected utility theory (to model the risk attitude of the decision maker);
2) Generalized Expected Utility Theory (to model the risk perception of the decision maker);
3) Multi-Attribute Utility Theory (MAUT) to model multi-criteria decision making;
The HDP incorporates a lifecycle decision making system, which includes risk aversion and risk perception of the decision maker(s) and stakeholder, and able to model the interactions between the humans, here represented from the building manager and the building occupants.