Informs 2015

60 / 552

INFORMS Philadelphia – 2015
58
SA58
58-Room 110A, CC
Multi-Agent Decision-Making for Smart Grids
Operation I
Sponsor: ENRE – Energy I – Electricity
Sponsored Session
Chair: Amin Kargarian, Carnegie Mellon University, 5000 Forbes Ave,
Pittuburgh, PA, 15232, United States of America,
1 - A Fast Distributed Algorithm for Power Optimization Problems
Ramtin Madani, Columbia University, 116th St. & Broadway,
New York, NY, 10027, United States of America,
, Abdulrahman Kalbat, Javad Lavaei
This talk aims to develop an efficient numerical algorithm for power optimization
problems by leveraging the small treewidth of real-world power grids. To this end,
a distributed algorithm is proposed for arbitrary conic programs with a low
treewidth, based on the alternating direction method of multipliers (ADMM). The
iterations of this algorithm consist of two subproblems, which are highly
parallelizable and enjoy closed-form solutions. The algorithm is tested on several
power systems.
2 - Demand Response using Supply Function Bidding
Na Li, Assistant Professor, Harvard University, 33 Oxford St,
MD 147, Cambridge, Ma, 02138, United States of America,
, Lijun Chen, Munther Dahleh
An abstract market model based on supply function bidding is proposed for
demand response to match power supply fluctuations. We characterize the
resulting equilibria in competitive and oligopolistic markets and analyze the
efficiency of the equilibria. The equilibrium in competitive market maximizes
social welfare, and the equilibrium in oligopolistic market has bounded efficiency
loss under certain mild assumptions. We also propose distributed algorithms to
achieve the equilibria.
3 - Community Storage for Firming
Chenye Wu, Postdoctoral Fellow, UC Berkeley, 60 Panoramic
Way, Berkeley, CA, 94704, United States of America,
, Kameshwar Poolla
We analyze the benefit of storage capacity sharing for a set of residential
consumers in a community. Each consumer has its own choice of either installing
its own storage system or investing in a shared storage system. In the latter case,
they must also decide on a scheme to allocate the costs. We formulate the
problem as a cooperative game and identify an efficient and stable cost allocation
rule. We further show this cost allocation rule induces a weakly incentive
compatible mechanism.
4 - Fully Distributed Approach for Optimal Power Flow Calculations
Javad Mohammadi, Carnegie Mellon University, Electrical and
Computer Engineering, CMU, Pittuburgh, PA, 15232,
United States of America,
In this talk, we propose a method which enables a fully distributed solution of the
DC Optimal Power Flow problem. The approach consists of an iterative procedure
that aims at solving the first order optimality conditions in fully distributed
manner at both nodal and regional level.
SA59
59-Room 110B, CC
Open Pit Mining
Sponsor: ENRE – Natural Resources I – Mining
Sponsored Session
Chair: Alexandra Newman, Professor, Colorado School of Mines,
Mechanical Engineering, Golden, CO, 80401, United States of America,
1 - Open Pit Mine Scheduling with Variants on
Inventory Considerations
Mojtaba Rezakhah, PhD Student, Colorado School of Mines,
Godlen, CO, 80401, United States of America,
We present several ways of modeling stockpiling (with and without considering
degradation) in open pit mine production scheduling, including (i) individual
stockpiles for each block and (ii) binned stockpiles with pessimistic grade
estimates. These models are formulated for a currently operational mine and
compared to results without stockpiling in order to assess the benefits of
stockpiling and to analyze the relationship between milling capacity and
stockpiling value.
2 - Models for Inventory Allocation of Erratic-demand Spare Parts in
a Multi-echelon System
Andrea Arias, PhD Student, Universidad Catolica de Valparaiso,
Avda Brasil 2241, Piso 6, Valparaiso, V, 2362807, Chile,
Jimena Pascual, Timothy Matis
Mining industry has processes that are carried out by complex machines with
many parts, which are reparable upon failure. We consider a two-level supply
chain inventory system for high cost/low demand repairable items. The aim is to
determine inventory levels for each item at every warehouse in order to
maximize system availability. We will present a comparative study between two
approaches used for addressing this problem from which some interesting ideas
for further research are proposed.
3 - Operational Flexibility in the Gold Mining Industry
Panos Markou, IE Business School, Calle Maria de Moina 12
Bajo, Madrid, 28006, Spain,
Daniel Corsten
Gold mining companies are highly exposed to commodity price risks. We examine
the effects of using “high-grading” (to combat these risks) on the operational and
financial performance of the miners. We find that although high-grading can
increase gross margin and profitability, it also increases profit variance and
inventory levels. Further, we show that these undesirable effects can be reduced
through financial hedging strategies.
4 - Using The Bienstock-Zuckerberg Algorithm for Large-scale,
Real-world Mine Planning Problems
Alexandra Newman, Professor, Colorado School of Mines,
Mechanical Engineering, Golden, CO, 80401, United States of
America, , Marcos Goycoolea,
Eduardo Moreno, Daniel Espinoza, Andrea Brickey
We apply an algorithm that solves the LP relaxation of precedence-constrained
knapsack problems to mine planning models which determine when to extract a
notional three-dimensional block of ore or waste, or when to commence an
activity (e.g., development, extraction) so as to maximize net present value,
subject to spatial precedence constraints and resource bounds. A simple heuristic
provides good integer-feasible solutions. Real-world instances from various
settings provide compelling results.
SA60
60-Room 111A, CC
Evaluating Student Learning
Sponsor: INFORM-ED
Sponsored Session
Chair: Sadan Kulturel-Konak, Professor, Pennsylvania State University,
Berks Campus, Reading, PA, 19609, United States of America,
1 - Assessing Students’ Global Awareness
Sadan Kulturel-Konak, Professor, Pennsylvania State University,
Berks Campus, Reading, PA, 19609, United States of America,
Students are expected to become increasingly globally aware in order to be better
prepared for a career in an international knowledge-based society. We define
global awareness knowledge, skills and abilities (KSA) that need to be met for a
student to be proficient in global awareness and define an assessment framework
based on the Model of Domain Learning (MDL). The preliminary findings for the
effectiveness of the proposed global awareness interest assessment will be
presented.
2 - A Web-based Peer Evaluation Tool for Professional
Skills Assessment
Abdullah Konak, Professor, Penn State Berks, Tulpehocken Road,
P.O. Box 7009, Reading, PA, 19610, United States of America,
We introduce the Peer Evaluation & Assessment Resource (PEAR), which is a
web-based solution that was created to efficiently assess the teamwork skills of
students through peer and self-evaluations. The PEAR application allows
instructors to form teams and choose a rubric to assess their teamwork skills and
contributions. In addition to built-in PEAR rubrics, instructors can also create
custom rubrics based on the Model of Domain Learning to better suit their specific
courses.
SA58