Presentation is loading. Please wait.

Presentation is loading. Please wait.

Context-Aware Parameter Estimation for Forecast Models in the Energy Domain Lars Dannecker 1,2, Robert Schulze 1, Matthias Böhm 2, Wolfgang Lehner 2, Gregor.

Similar presentations


Presentation on theme: "Context-Aware Parameter Estimation for Forecast Models in the Energy Domain Lars Dannecker 1,2, Robert Schulze 1, Matthias Böhm 2, Wolfgang Lehner 2, Gregor."— Presentation transcript:

1 Context-Aware Parameter Estimation for Forecast Models in the Energy Domain Lars Dannecker 1,2, Robert Schulze 1, Matthias Böhm 2, Wolfgang Lehner 2, Gregor Hackenbroich 1 1 SAP Research Dresden, 2 Technische Universität Dresden

2 ©2011 SAP AG. All rights reserved.2 Agenda 1. Forecasting in the Energy Domain 2. Context-Aware Forecast Model Repository 3. Experimental Evaluation 4. Summary and Future Work

3 Forecasting in the Energy Domain

4 ©2011 SAP AG. All rights reserved.4 Forecasting Process and Characteristics Predicting the Future  Quantitative model describing historic time series behavior  Uses parameters to represent specific characteristic  Estimated model mathematically calculates future behavior Specific Characteristics… …for energy time series Multi-Seasonality Dependence on external influences Evolving over time Negligible linear trend Continuous stream of measurements Base Component Trend Component Season Component ε ε

5 ©2011 SAP AG. All rights reserved.5 European Energy Market Market Organizer TSO BG2BG3 Supply Demand Balancing ForecastingAggregation BG1 Balancing Energy Demand and Supply Guarantee stable grids  Energy Demand has to be satisfied  Penalties for oversupply  Day-Ahead & intraday market  Integration of more RES in power mix  Accurate predictions at any point in time Renewable Energy Sources (RES) Increasing support  Depending on uncertain influences  Not plannable like traditional power  Accurate prediction for next day RES supply necessary

6 ©2011 SAP AG. All rights reserved.6 Energy Data Management for Evolving Time Series Energy Data Management Analytics close to the data  Quick reactions to changing time series  Always up-to-date forecasts Appending new values over time  Optimal parameters change and reoccur over time  Multiple local minima in parameter space  Continuous forecast model evaluation  Efficient forecast model adaptation

7 Context-Aware Forecast Model Repository

8 ©2011 SAP AG. All rights reserved.8 Context of Energy Time Series Influences for Supply and Demand  Time series development influenced by background processes  Changing context causes changes demand and supply behavior  Calendar: Special Days, Season  Meteorological: Wind speed, Temp.  Economical: Population Context Drift Different types of drifting context

9 ©2011 SAP AG. All rights reserved.9 Case-Based Reasoning = Learning how to solve new problems from past experience Energy domain: Seasonal reoccurring contexts  Reuse previous forecast models  Retain: Save old parameter combinations with their respective context  Retrieve: Search repository for a context most similar to the current context  Revise: Use parameter combinations of similar context as input for optimization Basic Idea Problem-Solution Case Base Start Values Updating trigger Continuous InsertionsContinuous Forecasts Updated Parameters Time series Current Forecast Model 3. Parameter Re-Estimation Global SearchLocal Search Forecast Error Calculation Retrieve Retain Starting Values for Estimation 2. Parameter Storing and Retrieval 1. Model Evaluation Model History Tree Insert Retrieve Distance Compuation ReviseRetain Revise Retrieve

10 ©2011 SAP AG. All rights reserved.10 Parameter Insertion ContextSummaryParametersEnd Index dayhouryearmeantemperaturep1p1 p2p2 K C1C C2C C3C C4C day temperature ≥ 7<7<7 < 11.3 hour <10≥ 10 ≥11.3 <8 hour ≥ 8 year <2004 ≥ 2004 year <2005≥ 2005 mean <500 ≥ 500 ContextSummaryParametersEnd Index dayhouryearmeantemperaturep1p1 p2p2 K C1C C2C C3C C4C C ContextSummaryParametersEnd Index dayhouryearmeantemperaturep1p1 p2p2 K C1C C2C C3C C4C C PIQR ContextSummaryParametersEnd Index dayhouryearmeantemperaturep1p1 p2p2 K C1C C2C ContextSummaryParametersEnd Index dayhouryearmeantemperaturep1p1 p2p2 K C1C C2C C year <2008 ≥ Traverse to leaf node 2. Insert 4. Chose attribute with highest 5. Split Tree Structured Repository  Decision nodes: Splitting attribute, splitting value  Leaf nodes: Set of parameter combinations, end index  Splitting attributes chosen using Partial Interquartil Range (PIQR)  Split via partitioning median 3.  True/Split

11 ©2011 SAP AG. All rights reserved.11 Parameter Retrieval x A B D E F G H I J K O P Q R T U N M O=(0.9,0.25) P=(0.85,0.2) R=(0.75,0.95) Q=(0.85,0.55) U=(0.95,0.85) T=(0.93,0.75) 1. Traverse to corresponding leaf node Find R as nearest neighbour Find O as nearest neighbor Find P as nearest neighbor A BD E FG H IJ KM N Bob-test with  False  Ascent 3. Bob-test with cyclical  True  Descent 4. Bob-test with  False

12 ©2011 SAP AG. All rights reserved.12 Optimization Subsequence Similarity  Find parameters that are associated with most similar time series shape  Using Pearson Cross Correlation Coefficient Subsequent Parallel Optimization  Parallel local and global parameter optimization  Local: Nelder Mead; Global: Simulated Annealing  Results from local optimization directly used  Parallel global search to consider areas not covered  Global search continues after local search finished  Quick accuracy recovery + global coverage Current subsequenceOld subsequence 2Old subsequence 1

13 Experiments NOTE: (Delete this element) Sample of title slide image. See SAP Image Library for other available images.

14 ©2011 SAP AG. All rights reserved.14 Settings DataSets  UK National Grid: Aggregated Demand United Kingdom  MeRegio: MeRegio project data 86 single customer demand  NREL Wind: Aggregated data from US wind parks  CRES PPV: Single appliance photovoltaic supply Forecast Models  Triple Seasonal Exponential Smoothing (5 parameters)  EGRV multi-equation autoregressive model (up to 31 parameters Comparison Scenario  Time vs. Accuracy against 4 common approaches Error Metric  Symetric Mean Absolute Percentage Error (SMAPE) Plattform  AMD Athlon 4850e (2.5 GHz), 4GB RAM, Windows 7  Visual C Subsequent Parallel Optimization  Parallel local and global parameter optimization  Results from local optimization directly used  Parallel global search to consider areas not covered  Global search continues after local search finished  Quick accuracy recovery + global coverage

15 ©2011 SAP AG. All rights reserved.15 Results: Triple Seasonal Exponential Smoothing TS-Exponential Smoothing  Small number of parameters, quick to estimate  MHT quickly reaches good accuracy  Our method is not superior on all data sets  Large result divergence for other approaches  MHT overhead: Eval (100 models)  4 msec, models  0.6 sec

16 ©2011 SAP AG. All rights reserved.16 Results: EGRV Model (Energy Domain Specific) EGRV  Large number of parameters, hard to estimate  MHT achieved best results on all data sets  Difference between best and worst approach much larger  MHT better suited for more complex models  MHT overhead: Eval (100 models)  6 msec, models  1.1 sec

17 Summary and Future Work

18 ©2011 SAP AG. All rights reserved.18 Summary & Future Work Problem Evolving energy time series require efficient forecast model estimation Summary Time series context influences time series development Case-based reasoning approach Store previous forecast model parameters for reuse with similar contextual situation Tree organized Context-Aware Forecast Model Repository Retrieve parameter by comparing current context to past context Parameters serve as input for optimization approaches Future Work Evaluate accuracy for approach without subsequent optimization Order attributes in tree using information criterion Further parallelization

19 Thank You! Contact information: Lars Dannecker SAP Research Dresden

20 ©2011 SAP AG. All rights reserved.20 No part of this publication may be reproduced or transmitted in any form or for any purpose without the express permission of SAP AG. The information contained herein may be changed without prior notice. Some software products marketed by SAP AG and its distributors contain proprietary software components of other software vendors. Microsoft, Windows, Excel, Outlook, and PowerPoint are registered trademarks of Microsoft Corporation. IBM, DB2, DB2 Universal Database, System i, System i5, System p, System p5, System x, System z, System z10, System z9, z10, z9, iSeries, pSeries, xSeries, zSeries, eServer, z/VM, z/OS, i5/OS, S/390, OS/390, OS/400, AS/400, S/390 Parallel Enterprise Server, PowerVM, Power Architecture, POWER6+, POWER6, POWER5+, POWER5, POWER, OpenPower, PowerPC, BatchPipes, BladeCenter, System Storage, GPFS, HACMP, RETAIN, DB2 Connect, RACF, Redbooks, OS/2, Parallel Sysplex, MVS/ESA, AIX, Intelligent Miner, WebSphere, Netfinity, Tivoli and Informix are trademarks or registered trademarks of IBM Corporation. Linux is the registered trademark of Linus Torvalds in the U.S. and other countries. Adobe, the Adobe logo, Acrobat, PostScript, and Reader are either trademarks or registered trademarks of Adobe Systems Incorporated in the United States and/or other countries. Oracle is a registered trademark of Oracle Corporation. UNIX, X/Open, OSF/1, and Motif are registered trademarks of the Open Group. Citrix, ICA, Program Neighborhood, MetaFrame, WinFrame, VideoFrame, and MultiWin are trademarks or registered trademarks of Citrix Systems, Inc. HTML, XML, XHTML and W3C are trademarks or registered trademarks of W3C®, World Wide Web Consortium, Massachusetts Institute of Technology. Java is a registered trademark of Sun Microsystems, Inc. JavaScript is a registered trademark of Sun Microsystems, Inc., used under license for technology invented and implemented by Netscape. SAP, R/3, SAP NetWeaver, Duet, PartnerEdge, ByDesign, SAP BusinessObjects Explorer, StreamWork, and other SAP products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of SAP AG in Germany and other countries. © 2011 SAP AG. All rights reserved Business Objects and the Business Objects logo, BusinessObjects, Crystal Reports, Crystal Decisions, Web Intelligence, Xcelsius, and other Business Objects products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of Business Objects Software Ltd. Business Objects is an SAP company. Sybase and Adaptive Server, iAnywhere, Sybase 365, SQL Anywhere, and other Sybase products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of Sybase, Inc. Sybase is an SAP company. All other product and service names mentioned are the trademarks of their respective companies. Data contained in this document serves informational purposes only. National product specifications may vary. The information in this document is proprietary to SAP. No part of this document may be reproduced, copied, or transmitted in any form or for any purpose without the express prior written permission of SAP AG. This document is a preliminary version and not subject to your license agreement or any other agreement with SAP. This document contains only intended strategies, developments, and functionalities of the SAP® product and is not intended to be binding upon SAP to any particular course of business, product strategy, and/or development. Please note that this document is subject to change and may be changed by SAP at any time without notice. SAP assumes no responsibility for errors or omissions in this document. SAP does not warrant the accuracy or completeness of the information, text, graphics, links, or other items contained within this material. This document is provided without a warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability, fitness for a particular purpose, or non-infringement. SAP shall have no liability for damages of any kind including without limitation direct, special, indirect, or consequential damages that may result from the use of these materials. This limitation shall not apply in cases of intent or gross negligence. The statutory liability for personal injury and defective products is not affected. SAP has no control over the information that you may access through the use of hot links contained in these materials and does not endorse your use of third-party Web pages nor provide any warranty whatsoever relating to third-party Web pages.


Download ppt "Context-Aware Parameter Estimation for Forecast Models in the Energy Domain Lars Dannecker 1,2, Robert Schulze 1, Matthias Böhm 2, Wolfgang Lehner 2, Gregor."

Similar presentations


Ads by Google