Faculty Project Proposal Summary
Calendar Year 2014 Program
Faculty Name & Department/School: Dr. Meghna Babbar-Sebens, School of Civil and Construction Engineering
Research Project Title: Direct interaction in human-computer collaborative optimization of sustainable conservation practices in watersheds
Project Overview: This research investigates an innovative community-driven optimization approach for designing conservation practices in watersheds. The proposed method is built on a democratic design approach in which a watershed’s stakeholder community interacts via web-based technologies and visual interfaces and collaborates with each other and with the underlying data-driven models (e.g., environmental models, machine learning and computational optimization algorithms) to create alternatives of practices that are more acceptable and feasible on their landscape. In this manner, they explore solutions that not only work on their local area, but also, when coupled with practices adopted by other users, provide benefits to the entire watershed system. The overarching goal of this research will be to examine multiple methods for collaborative design and interactive optimization that allow the users to directly modify alternatives of conservation practices in the simulated landscape, and the effect of that direct interaction on performance of the search process.
What skills will students obtain in this project? Students in Computer Science with background in web development and information systems, and/or students interested in Water Resources with programming background are encouraged to apply. Student will learn multiple aspects of this interdisciplinary project:
Student research tasks: The successful candidate will investigate the following research questions related to direct user interaction in a human-computer collaborative optimization process:
1) What kind of visualization approaches and map API are needed to best facilitate direct user interaction for dynamic modification of spatial attributes and features of conservation practices in the alternative designs? For example, do users prefer to draw polygons/polylines on maps to suggest changes in physical design (i.e., best management practices), or is it better for them to indicate changes via text or numerical values?
2) What is the most effective way to include these user-generated alternatives in the optimization algorithm’s search process? For example, should they be used to replace low performing solutions in the optimization algorithm’s population or should they be added to the population probabilistically?
3) Does direct user interaction improve or worsen the optimization algorithm’s convergence rates? For example, when users directly modify alternatives does that change the optimization trajectory towards faster identification of optimal and acceptable solutions?
Number of hours per week expected of student: 10-15 hrs in three large blocks
Preferred Faculty Contact Information (email): Meghna@oregonstate.edu
Download student application HERE
Back to Main Page HERE