Select a calendar:
Filter May Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Events for May 01, 2008
-
Can the Earth Produce the Biomass We Demand
Thu, May 01, 2008 @ 12:45 PM
Mork Family Department of Chemical Engineering and Materials Science
Conferences, Lectures, & Seminars
Lyman Handy Colloquium presentsProfessor Tadeusz W. PatzekDepartment of Civil and Environmental Engineering
University of California - BerkeleyAbstractTo demonstrate the utter impossibility of sustained, industrial-scale production of biofuels anywhere and from and source, I consider the local,
field-scale sustainability of a productive industrial maize agrosystem that has replaced a fertile grassland ecosystem. Using the revised Second Law approach of Svirezhev, I show that currently this agrosystem is unsustainable in the US (and anywhere else), with or without tilling the
soil. The calculated average erosion rates of soil necessary to dissipate the entropy produced by US maize agriculture, 23 -- 45 t/ha-yr, are bounded
from above by an experimental estimate of mean soil erosion by conventional agriculture worldwide, 47 t/ha-yr. Between 1982 and 1997, US agriculture
caused an estimated 7 -- 23 t/ha-yr of average erosion with the mean of 15 t/ha-yr. The lower mean erosion rate of no till agriculture, 1.5 t/ha-yr,
necessitates the elimination of weeds and pests with field chemicals -- with the ensuing chemical and biological soil degradation, and chemical runoff -- to dissipate the produced entropy. The increased use of field chemicals that replace tillers is equivalent to the killing or injuring of up to 300 kg/ha-yr of soil flora and fauna. Additional soil degradation, not discussed
here, occurs by acidification, buildup of insoluble metal compounds, and buildup of toxic residues from field chemicals. The degree of unsustainability of an average US maize field is high, requiring 6 -- 13 times more energy to remediate soil degradation, etc., than the direct energy inputs to maize agriculture. This additional energy, if spent, would
not increase maize yields. The calculated ``critical yield'' of ``organic'' maize agriculture that does not use field chemicals and fossil fuels is only 30 percent lower than the average maize yield of 8.7 tons per hectare (140 bushel/acre) assumed here. Immediate attention should be devoted in the US to more sustainable alternatives to the current industrial agriculture. I will also discuss the implications of my findings to the current wholesale destruction of the tropical ecosystems. URL: http://petroleum.berkeley
Location: Olin Hall of Engineering (OHE) - 122
Audiences: Everyone Is Invited
Contact: Petra Pearce Sapir
-
System Earthquake Risk Assessment (SERA)
Thu, May 01, 2008 @ 02:00 PM - 03:00 PM
Sonny Astani Department of Civil and Environmental Engineering
Conferences, Lectures, & Seminars
Dr. Dennis Ostrom, Consultant - San Diego Gas & Electric, Pacific Gas & Electric, and Southern California EdisonDr. Ostrom will present the Development of a utility developed System Earthquake Risk Assessment (SERA) and present it from a utility perspective. Characterization of the hazard, component vulnerability, system vulnerability, development of performance goals, return to service times, system performance and future studies and/or research will be discussed and presented.
Location: Kaprielian Hall (KAP) - 209
Audiences: Everyone Is Invited
Contact: Evangeline Reyes
-
CS Colloq: AND/OR Search Strategies for Combinatorial Optimization in Graphical Models
Thu, May 01, 2008 @ 03:30 PM - 04:30 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Title: AND/OR Search Strategies for Combinatorial Optimization in Graphical ModelsSpeaker: Radu Marinescu (UCI)Abstract:
The AND/OR search space for graphical models is a new framework for search that is sensitive to the independencies in the model, often resulting in exponentially reduced complexities. The AND/OR search tree search is in most cases exponentially smaller (and never larger) than the OR search tree. The AND/OR search graph is exponential in the treewidth of the graph, while the OR search graph is exponential in the pathwidth. We introduce a new generation of depth-first Branch-and-Bound as well as best-first AND/OR search algorithms that explore the context minimal AND/OR graph for solving general constraint optimization problems over graphical models. In conjunction with the AND/OR search space we also investigate a class of partitioning-based heuristic functions, based on the Mini-Bucket approximation that was shown to be powerful for optimization problems in the context of OR search spaces. Since variable selection can have a dramatic impact on search performance, we also introduce a class of depth-first AND/OR Branch-and-Bound and best-first AND/OR search algorithms that can accommodate various dynamic variable ordering heuristics. An extensive empirical evaluation showed conclusively that the new AND/OR search approach improves considerably over the traditional OR search, on a variety of probabilistic and deterministic benchmarks. We next apply the AND/OR perspective to decision diagrams. We extend them with AND nodes capturing function structure decomposition, resulting in AND/OR Multi-Valued Decision Diagrams (AOMDDs). The AOMDD is a canonical form that compiles a graphical model and has size bounded exponentially by the treewidth, rather than pathwidth (as is the case for OR decision diagrams). We present an AND/OR search based algorithm for compiling AOMDDs, as representations of the optimal set of solutions of a constraint optimization problem. An extensive experimental evaluation proved the efficiency of the weighted AOMDD data structure.Location: Seaver Science Library (SSL) - 150
Audiences: Everyone Is Invited
Contact: CS Colloquia
