A Stochastic Approach to Model Household Re-occupancy in A Community Following A Natural Hazard

Abstract

additionally, based on available social science studies, we further define a set of time-dependent conditional re-occupancy probability functions (CRPFs) that give the probability of a household re-occupying its pre-event dwelling units at any time conditional on the change in households JFS and its income level. Finally, the time-variant household-level re-occupancy probability is derived by solving the DTMC with partially absorbing boundary conditions described by the CRPFs. The community-level HRO is then obtained through aggregating the household-level re-occupancy state across the community over the recovery time horizon. The model will be further calibrated by data collected in ongoing field studies with an ultimate goal of supporting further researches on community resilience planning.

The re-occupancy of displaced households in a community following a hazard event is a complex social process driven collectively by the functionality states of community building portfolios and supporting lifelines. This study presents a novel approach for household re-occupancy (HRO) modeling using discrete state, Discrete Time Markov Chain (DTMC). Our hypothesis is that the reoccupancy state of a displaced household at a post-event time is collectively determined by the joint functionality status (JFS) (of its related school(s), workplace(s) and pre-event dwelling unit) and by the resourcefulness of the household largely determined by its income level. Accordingly, we construct a one-step transition probability matrix of the DTMC modeling the households JFS as a function of the functionality states of its school(s), workplace(s) and pre-event dwelling unit