Degree Type

Thesis

Date of Award

2018

Degree Name

Master of Science

Department

Industrial and Manufacturing Systems Engineering

Major

Industrial and Manufacturing Systems Engineering

First Advisor

Caroline C. Krejci

Second Advisor

Michael C. Dorneich

Abstract

This thesis consists of one submitted journal paper and one modified conference paper, both on simulating residents' weatherization adoption decisions.

Residential buildings are responsible for a large number of energy consumptions and are therefore a major contributor to climate change. Weatherization is set of approaches that can be used to make buildings more energy-efficient. This yields many benefits for residents, including reduced energy costs and improved health and safety, as well as benefits for the environment, such as reducing greenhouse gas emissions and conserving resources. However, the current adoption rate for such a good practice remains low. The government has been trying to encourage the adoptions, but its efforts have had limited success. Therefore, it is of great practical significance to explore residents' weatherization adoption decisions, and then assist stakeholders in encouraging weatherization.

This thesis proposes two hybrid simulation models of an urban neighborhood populated with autonomous "households" for their weatherization adoption decisions. They both include three modeling techniques: 1) a building energy simulation model, 2) an agent-based model, 3) a social network model. The building energy simulation model is used to calculate monthly energy consumption data of residential buildings under pre- and post- weatherized conditions, which inform households of the potential energy savings that would result from the decisions to weatherize homes. The agent-based model is used to model the detailed decision-making process of households and peer interactions among them about weatherization. The realistic topology of households' social environment, in which peer interactions and information diffusion take place, is captured by the social network model. The hybrid models represent the complex dynamic feedback loop that connects households' weatherization decisions, energy-related decision outcomes, and communication of outcomes among community members, all of which influence future decisions.

The first hybrid model models a block of 29 households and their physical social networks (i.e., the physical neighborhood). The effects of households' and communities' internal attributes on households' weatherization adoption decisions are explored. Experimental results suggest that more households tend to weatherize when there is a self-weatherized leader or they have short memories about energy bills, and especially in a denser social network as it is a self-reinforcing circle where positive information about weatherization can spread more widely. Four corresponding recommended policies to improve the adoptions of weatherization are also discussed.

The second hybrid model develops an agent-based model with household agents and media agents. It is embedded in a multilayer social network, which allows households to interact via both a physical social network (i.e., their neighborhood) and a virtual social network (i.e., online). We evaluate the strength of social interactions based on households' local centrality, spatial location, and social demographics. Opinion dynamics of households are captured by the Susceptible-Exposed-Infected-Recovered epidemic model. This model is used to explore the effects of more characteristics of households, media, and communities, and evaluate different interventions that government and other policy makers could adopt in an effort to promote weatherization adoption in residential buildings. Experimental results demonstrate the necessity of modeling a multilayer social network and the slight usefulness of a higher value of randomness of the physical social network. The results also indicate that selecting the most important (influential) households as early adopters of weatherization could significantly promote the weatherization adoptions. In addition, slightly increasing the probability of an approved applicant being served by WAP (e.g., the current real-life value in Iowa is 2.5%) yields dramatic increases in weatherization adoptions. How to greatly leverage the effects of media on households' weatherization adoption decisions is also discussed.

Copyright Owner

Wanyu Huang

Language

en

File Format

application/pdf

File Size

105 pages

Share

COinS