Computational Sustainability Carla P. Gomes Institute for - PowerPoint PPT Presentation

Computational Sustainability Carla P. Gomes Institute for Computational Sustainability Computing and Information Science Cornell University Support : Role of Information Sciences and Engineering in Sustainability Expeditions RISES in Computing (CISE) DC, Feb. 2011

The Institute for Computational Sustainability (ICS) Research Team ICS members and collaborators for their many contributions towards the development of a vision for research, education, and outreach activities in the new Expeditions area of Computational Sustainability in Computing (CISE) Support : 29 graduate students 24 undergrad. students

Sustainability and Sustainable Development The 1987 UN report, � Our Common Future � (Brundtland Report): ! ! Raised serious concerns about the State of the Planet. ! ! Introduced the notion of sustainability and sustainable development: Sustainable Development: � development that meets the needs of the present without compromising the ability of future generations to meet their needs. � Gro Brundtland UN World Commission on Environment and Development,1987. Norwegian Prime Minister Chair of WCED 3

Follow-Up Reports: Intergovernmental Panel on Climate Change (IPCC 07) Global Environment Outlook Report (GEO 07) � There are no major issues raised in Our Common Future for which the foreseeable trends are favourable. � Erosion of Biodiversity Examples: • ! At the current rates of human destruction of natural ecosystems, 50% of all species of life on earth will be extinct in 100 years. Wilson, E.O. The Future of Life (2002) . Global Warming • ! The biomass of fish is estimated to be 1/10 of what it was 50 years ago and is declining. Worm et al. (2006) +130 [Nobel Prize with Gore 2007] 4 countries

Safe Operating Boundaries: Crucial Biophysical Systems Source: Planetary Boundaries: A Safe Operating Space for Humanity , Nature, 2009 5

Sustainability: Interlinked environment, economic, and social issues Our Common Future recognized that environmental, economic and social issues are interlinked. " The economy only exists in the context of societies, and both society and economic activity are constrained by the earth � s natural systems. " A secure future depends upon the health of all 3 systems (environment, society, economy). Sustainable Development encompasses balancing environmental, economic, and societal needs. 6

Challenging Sustainability Problems Complex Dynamical Systems Sustainability problems are unique in scale, impact, complexity, and richness: Often involving multiple and highly interconnected ecosystems.noaa.gov components and players, in highly dynamic and Natural Ecosystem uncertain environments. Offer challenges but also opportunities for the advancement of the state of the art in computing and information science. Smart Grid: Complex Digital Ecosystem 7

Computational Sustainability: Vision Computer scientists can — and should — play a key role in increasing the efficiency and effectiveness of the way we manage and allocate our natural resources, while enriching and transforming Computing and Information Science and related disciplines. We need critical mass in the new field of Computational Sustainability!!! 8

Outline I Examples of Computational Sustainability problems highlighting research themes II Our research themes III Building a community in Computational Sustainability IV Conclusions 9

Sample of Computational Sustainability Problems I Environment E.g. Conservation and Biodiversity II Socio-Economic Systems E.g. Poverty mapping and poverty reduction, and harvesting policies. III Energy Fuel distributors Gasoline E.g. Smart grid, Material discovery and Biofuels Farmers Non-energy producers crops Food Energy Consumers supply crops Energy Water quality market Environmental Soil quality Social welfare impact Biodiversity Economic Local air 10 pollution impact

Conservation and Biodiversity 11

Combatting Biodiversity Loss: Landscape Connectivity Ideas from circuit theory for mapping critical linkages in complex landscapes Brad McRae et al. Connectivity for mountain in southern California. Maintaining movement and connectivity across Blue - low current density (low densities of dispersing mountain lions); landscapes can help reduce inbreeding, Yellow - movement bottlenecks, where connectivity is most vulnerable to hab destruction. increase genetic diversity, Red - high current flow high priority areas for conservation or restoration. and/or colonize new habitat. � With limited funding and constant threats of habitat loss how do we choose which habitats to protect so that landscapes will stay well-connected for wild animal species? � Opportunities for new computational models integrating ecological and economic constraints.

Combatting Habit Loss and Fragmentation: Wildlife Corridors Wildlife Corridors link core biological areas, allowing animal movement between areas. Typically: low budgets to implement corridors. Example : Goal: preserve grizzly bear populations in the U.S. Northern Rockies by creating wildlife corridors connecting 3 reserves: Yellowstone National Park Glacier Park / Northern Continental Divide cost suitability Salmon-Selway Ecosystem

Conservation and Biodiversity: Wildlife Corridors Challenges in Constraint Reasoning and Optimization Wildlife corridor design Computational problem " Connection Sub-graph Problem Connection Sub-graph Problem Given a graph G with a set of reserves: Find a sub-graph of G that: contains the reserves; Connection Sub-Graph - NP-Hard is fully connected; with cost below a given budget; Worst Case Result --- Real-world problems possess and hidden structure that can be exploited allowing with maximum utility scaling up of solutions. Interdisciplinary Research Project: Conrad, Dilkina, Gomes, van Hoeve, Sabharwal, Sutter; 2007-2010 14

Minimum Cost Corridor for the Connected Sub-Graph Problem ! ! !"#$%&'()*+)&,' " ' !"#$%&'($)(*"#*+$,+** '-%$.*&/'' ! ! 0+1&2'34%4/&5&%'5%4654.*&'7'3$*8#$/+4*')/&',$*94.*&':$%';1&2'<,/4**=' #(/.&%'$:'5&%/+#4*,'$%'%&,&%9&,'<#$5'(,&2'+#'6$#,&%94)$#'*+5&%45(%&=' 50x50 grid 40x40 grid 25x25 grid 10x10 grid 25 km 2 hex 25 km 2 hex 167 Cells 242 Cells 570 Cells 3299 Cells 1288 Cells Extend with $1.3B $891M $449M $99M $7.3M 2xB=$15M <1 sec <1 sec <1 sec 10 mins 2 hrs 10x in Util What if we were allowed extra budget? Need to solve problems large number of cells! " >64*4.+*+58'!,,(&,''

Models Are Important!!! Single Commodity Flow Quite compact (poly size) Directed Steiner Tree Exponential Number of Constraints $ ! Captures Better the Connectedness Structure # ! Provides good upper bounds (discover good cuts with ML) $ ! Dilkina and Gomes 2010

Science of Computation - Understanding Structure: � Typical � Case Analysis and Identification of Critical Parameters (Synthetic Instances) Runtime How is hardness affected as the budget fraction is varied? Problem evaluated on semi-structured graphs 3 reserves m x m lattice / grid graph with k terminals Inspired by the conservation corridors problem Place a terminal each on top-left and bottom-right Maximizes grid use Place remaining terminals randomly Assign uniform random costs and utilities from {0, 1, ! , 10} Runtime for Optimal Solution Utility Gap (Optimally Extended Min cost/ Optimal) No reserves: � pure optimization � � � 3 reserves From 6x6 to 10x10 grid (100 parcels): 1000 instances per data-point;

Real world instance: Glacier Park Corridor for grizzly bears in the Northern Rockies, connecting: Yellowstone Salmon-Selway Ecosystem Glacier Park Salmon-Selway (12788 nodes) Yellowstone Scaling up Solutions by Exploiting Structure 5 km grid 5 km grid (12788 land parcels): • ! Synthetic generator / Typical Case Analysis (12788 land parcels): minimum cost solution +1% of min. cost • ! Identification of Tractable Sub-problems • ! Exploiting structure via Decomposition Static/Dynamic Pruning • ! Streamlining for Optimization • ! New Encodings Our approach allows us to handle large problems and reduce corridor cost dramatically (hundreds of millions of dollars) compared to existing approaches while providing guarantees of optimality in terms of utility: Optimal or within 1% of optimality for interesting budget levels.

Multiple Species Identification of new problems to address multiple species: E.g. Steiner Multigraph Problem Wolverines Upgrading Shortest Path Problem Dynamics and Game Theory : • ! Study dynamics of interactions Lynx Grizzly Bear • ! How to be fair??? • ! What metrics? Collaborators: Michael K. Schwartz USDA Forest Service, Rocky Mountain Research Station Claire Montgomery Oregon State University 19

Outreach and Education Pedagogical Games Shortest path, Steiner trees, and much more about Computational Sustainability Boynton Middle School Math Day � Edutainment � � Video Games for middle school Lots of undergrad/Meng students Having fun designing games for A Travel Museum on Computational Effort led by Sustainability David 20 Schneider