Data requirements for stochastic solvers H.I. Gassmann Dalhousie University Halifax, Canada.

Slides:



Advertisements
Similar presentations
Pattern Recognition and Machine Learning
Advertisements

Bayesian Belief Propagation
Statistical perturbation theory for spectral clustering Harrachov, 2007 A. Spence and Z. Stoyanov.
Sample Approximation Methods for Stochastic Program Jerry Shen Zeliha Akca March 3, 2005.
1 Finite Constraint Domains. 2 u Constraint satisfaction problems (CSP) u A backtracking solver u Node and arc consistency u Bounds consistency u Generalized.
Solver and modelling support for stochastic programming H.I. Gassmann, Dalhousie University Happy Birthday András, November 2009.
Approximation Algorithms Chapter 14: Rounding Applied to Set Cover.
Computational Stochastic Optimization:
Nonlinear Programming
Randomized Algorithms Randomized Algorithms CS648 Lecture 20 Probabilistic Method (part 1) Lecture 20 Probabilistic Method (part 1) 1.
Applications of Stochastic Programming in the Energy Industry Chonawee Supatgiat Research Group Enron Corp. INFORMS Houston Chapter Meeting August 2, 2001.
1 Fast Primal-Dual Strategies for MRF Optimization (Fast PD) Robot Perception Lab Taha Hamedani Aug 2014.
Decision Making: An Introduction 1. 2 Decision Making Decision Making is a process of choosing among two or more alternative courses of action for the.
Principal Component Analysis
Stochastic Differentiation Lecture 3 Leonidas Sakalauskas Institute of Mathematics and Informatics Vilnius, Lithuania EURO Working Group on Continuous.
Introduction to Linear and Integer Programming Lecture 7: Feb 1.
Math443/543 Mathematical Modeling and Optimization
CPSC 322, Lecture 17Slide 1 Planning: Representation and Forward Search Computer Science cpsc322, Lecture 17 (Textbook Chpt 8.1 (Skip )- 8.2) February,
Integer Programming Difference from linear programming –Variables x i must take on integral values, not real values Lots of interesting problems can be.
Chapter 14 Simulation. Monte Carlo Process Statistical Analysis of Simulation Results Verification of the Simulation Model Computer Simulation with Excel.
Distributed Combinatorial Optimization
1 Wavelet synopses with Error Guarantees Minos Garofalakis Phillip B. Gibbons Information Sciences Research Center Bell Labs, Lucent Technologies Murray.
CSE 221: Probabilistic Analysis of Computer Systems Topics covered: Discrete time Markov chains (Sec. 7.1)
Daniel Kroening and Ofer Strichman Decision Procedures An Algorithmic Point of View Deciding ILPs with Branch & Bound ILP References: ‘Integer Programming’
Collaborative Filtering Matrix Factorization Approach
Decision Procedures An Algorithmic Point of View
June 2001 Universita’ degli Studi di Bergamo Corso di dottorato di ricerca 1 Asset/Liability Management Models in Insurance.
Approximation Algorithms for Stochastic Combinatorial Optimization Part I: Multistage problems Anupam Gupta Carnegie Mellon University.
Mathematical Programming Models for Asset and Liability Management Katharina Schwaiger, Cormac Lucas and Gautam Mitra, CARISMA, Brunel University West.
Recent Changes to the Optimization Services Project H.I. Gassmann, Faculty of Management J. Ma, Northwestern University R.K. Martin, The University of.
Nonlinear Data Assimilation and Particle Filters
Applied stochastic programming models and computation H.I. Gassmann Dalhousie University.
New developments in OSiL, OSoL and OS H.I. Gassmann R.K. Martin,J. Ma INFORMS Annual meeting, Washington, DC, October 2008.
17 July 2006DIMACS 2006 COIN-SMI 2006 Alan King IBM TJ Watson Research Center.
Stochastic Linear Programming by Series of Monte-Carlo Estimators Leonidas SAKALAUSKAS Institute of Mathematics&Informatics Vilnius, Lithuania
Zorica Stanimirović Faculty of Mathematics, University of Belgrade
Module 1: Statistical Issues in Micro simulation Paul Sousa.
Design of an Evolutionary Algorithm M&F, ch. 7 why I like this textbook and what I don’t like about it!
PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 3: LINEAR MODELS FOR REGRESSION.
Approximation Algorithms for Stochastic Optimization Chaitanya Swamy Caltech and U. Waterloo Joint work with David Shmoys Cornell University.
Bekkjarvik, A Heuristic Solution Method for a Stochastic Vehicle Routing Problem Lars Magnus Hvattum.
An introduction to stochastic programming H.I. Gassmann.
© 2015 McGraw-Hill Education. All rights reserved. Chapter 19 Markov Decision Processes.
An XML-based schema for stochastic programs H.I. Gassmann, R. Fourer, J. Ma, R.K. Martin EURO XXI, June 2006, Reykjavik.
Robust Optimization and Applications Laurent El Ghaoui IMA Tutorial, March 11, 2003.
Multi-state System (MSS) Basic Concepts MSS is able to perform its task with partial performance “all or nothing” type of failure criterion cannot be.
Quality of LP-based Approximations for Highly Combinatorial Problems Lucian Leahu and Carla Gomes Computer Science Department Cornell University.
OSiL: An XML-based schema for stochastic programs H.I. Gassmann, R. Fourer, J. Ma, R.K. Martin SP XI Vienna, August 2007.
By: Jesse Ehlert Dustin Wells Li Zhang Iterative Aggregation/Disaggregation(IAD)
Linear Programming Chapter 1 Introduction.
Optimally Modifying Software for Safety and Functionality Sampath Kannan U.Penn (with Arvind Easwaran & Insup Lee)
MS&E 348 The L-Shaped Method: Theory and Example 2/5/04 Lecture.
National Taiwan University Department of Computer Science and Information Engineering An Approximation Algorithm for Haplotype Inference by Maximum Parsimony.
CS4234 Optimiz(s)ation Algorithms L2 – Linear Programming.
6.5 Stochastic Prog. and Benders’ decomposition
Decomposition Methods
Chapter 9 Integer Programming
Collaborative Filtering Matrix Factorization Approach
Stochastic Network Interdiction
Lecture 2 – Monte Carlo method in finance
Dimension reduction : PCA and Clustering
Flow Feasibility Problems
Mechanical Engineering Department
An XML-based schema for stochastic programs
Representation Probabilistic Graphical Models Local Structure Overview.
EM based Multiuser detection in Fading Multipath Environments
6.5 Stochastic Prog. and Benders’ decomposition
Discrete Optimization
Presentation transcript:

Data requirements for stochastic solvers H.I. Gassmann Dalhousie University Halifax, Canada

Overview b Introduction b Algorithms b Data structures b Conclusions

Multistage stochastic linear program Any data item with nonzero subscript may be random (including dimensions where mathematically sensible) ~ stands for arbitrary relation ( , =,  )

Constraints involving random elements  means ~ with probability 1 or with probability at least  or with expected violation at most v

Problem classes b Recourse problems All constraints hold with probability 1All constraints hold with probability 1 b Chance-constrained problems Typically single stageTypically single stage b Hybrid problems Recourse problems including probabilistic constraints (VaR) or integrated chance constraints (CVaR)Recourse problems including probabilistic constraints (VaR) or integrated chance constraints (CVaR) Regulatory necessityRegulatory necessity Often modelled using integer variables and/or linking constraintsOften modelled using integer variables and/or linking constraints

Algorithms b Direct solution of the deterministic equivalent “Curse of dimensionality”“Curse of dimensionality” b Decomposition Recognize structureRecognize structure Repeated calls to solver with different dataRepeated calls to solver with different data Sampling of scenarios during algorithmSampling of scenarios during algorithm –stochastic decomposition –successive approximation –“EVPI relaxation” Scenario generator between AML and solverScenario generator between AML and solver

Data Structures b Often O(10 6 ) variables and constraints b Most compact representation possible Packed matrix format is insufficientPacked matrix format is insufficient Blocks corresponding to nodes in the event treeBlocks corresponding to nodes in the event tree Change blocks if problem dimensions are deterministicChange blocks if problem dimensions are deterministic A stj = A st0 +  A stj (addition or replacement)A stj = A st0 +  A stj (addition or replacement)

Conclusions b Stochastic extensions are difficult b Time is right