Events Calendar

Oral Dissertation Defense

Speaker: Maud Comboul, CEE Ph.D, Candidate

Talk Title: Stochastic and Multiscale Models for Urban and Natural Ecology

Abstract:
This research reflects on both particular cases of ecologies: sensor networks in urban water distribution systems and forest dynamics under changing disturbance regime, and the elaboration of exploration tools to investigate and assess those complex ecologies. The predictive models that emerged from this work rely on sophisticated computer simulations designed based on stochastic and multiscale principles, and contribute to characterizing and evaluating the knowledge and information that is available about the systems under investigation. This thesis is organized into three distinct projects, each of which confronts specific challenges encountered when modeling complex systems. The first project considers the monitoring of water distribution networks where we describe a stochastic parameterization and analysis of uncertainty for the design of single-stage as well as two-stage sensor networks aimed at maximizing the probability of detection of accidents and intrusions in water distribution systems. Next, this study explores the ecological and evolutionary impacts of different disturbance regimes (generated following a stochastic Poisson process) on forests using the framework of a spatially explicit and individual-based forest model designed around four functional traits of trees. The final project investigates a multiscale characterization of forest dynamics using Monte-Carlo simulations of the fine scale dynamics to synthesize a coarse-grain stochastic model describing the dynamics of the system on larger spatial scales. In addition to modeling aspects, all three projects yield adequate computational performance, thereby assessing a recurrent challenge associated with the computational feasibility and performance of relevant numerical algorithms.

Advisor: Dr. Roger Ghanem

Location: Kaprielian Hall (KAP) - 209

Audiences: Everyone Is Invited

Contact: Evangeline Reyes