a project by Approach I: Model-based computing time reduction - - PowerPoint PPT Presentation

▶
a project by approach i model based computing time
SMART_READER_LITE
LIVE PREVIEW

a project by Approach I: Model-based computing time reduction - - PowerPoint PPT Presentation

Nov 28, 2022 360 likes •595 views

Methods to reduce computing times of linear energy system optimization models IAEE 2019, Ljubljana Yvonne Scholz , Karl Kin Cao, Manuel Wetzel, Kai von Krbek DLR German Aerospace Center, Department of Energy Systems Analysis Daniel

slide-1
SLIDE 1

a project by

Yvonne Scholz, Karl Kiên Cao, Manuel Wetzel, Kai von Krbek DLR – German Aerospace Center, Department of Energy Systems Analysis Daniel Rehfeldt, Thorsten Koch Zuse Institut Berlin / TU Berlin

IAEE 2019, Ljubljana

Methods to reduce computing times of linear energy system

  • ptimization models
slide-2
SLIDE 2

Approach I: Model-based computing time reduction

Speed-Up strategies Solver-based Solver parameters Solving methodology Model-based Pure model reduction Heuristic decomposition Exact Decomposition

slide-3
SLIDE 3

Speed-Up strategies Solver-based Solver parameters Solving methodology Model-based Pure model reduction Heuristic decomposition Exact Decomposition

„Low Hanging Fruits“

slide-4
SLIDE 4

4

  • Selection of measures (also useful to decrease memory need):

– Input data should not differ much in its order of magnitude – Index order influences computing time

  • Useful, but not necessarily faster
  • Assignment statements with a different set order can be faster
  • It can be better to place large index sets at the beginning

– Use of “option kill” , e.g. for long time-series input parameters saves memory – Abundant use of “Dollar Control over the Domain of Definition” – Consistent (and limited) use of defined variables – Avoid the consideration of technologies providing the same service at the same costs – Consider alternative formulation of model constraints (dense vs. sparse)

  • Helpful references: “Speeding up GAMS Execution Time”

by Bruce A. McCarl https://www.gams.com/mccarl/speed.pdf

Source code improvement

Yvonne Scholz (DLR)

slide-5
SLIDE 5

Speed-Up strategies Solver-based Solver parameters Solving methodology Model-based Pure model reduction Heuristic Decomposition Exact Decomposition

Approach I: Model-based computing time reduction

slide-6
SLIDE 6

Presented speed-up approaches

Heuristic decomposition Rolling time horizons Myopic technology expansion planning “Spatial zooming“ “Temporal zooming” Increasing technological detail Pure model reduction Slicing Representative time intervals Focusing regions of interest Neglecting technologies Aggregation Temporal downsampling Spatial downsampling (building network equivalents) Defining technology classes

slide-7
SLIDE 7

Evaluation methodology

slide-8
SLIDE 8

8

Model name REMix Author (Institution) German Aerospace Center (DLR) Model type Linear programming minimization of total system costs economic dispatch / optimal dc power flow with expansion of storage and transmission capacities Sectoral focus Electricity Geographical focus Germany Spatial resolution 488 nodes Analyzed year (scenario) 2030 Temporal resolution 8760 time steps (hourly)

Evaluation: Overview

Solver Commercial Algorithm Barrier Cross-over Disabled

  • Max. parallel

barrier threads 16 Scaling Aggressive Yvonne Scholz (DLR)

slide-9
SLIDE 9

Results

slide-10
SLIDE 10

10

Performance

Results: Spatial aggregation

Yvonne Scholz (DLR)

slide-11
SLIDE 11

11

Performance Accuracy

Results: Spatial aggregation

Yvonne Scholz (DLR)

slide-12
SLIDE 12

12

Performance Accuracy

  • Speed-up factor: ≈5
  • Accuracy error mainly < 10 % (grids: ≈20%)

Results: Spatial aggregation

Yvonne Scholz (DLR)

slide-13
SLIDE 13

13

Results: Temporal zooming

Yvonne Scholz (DLR)

Performance Accuracy

slide-14
SLIDE 14

14

Performance Accuracy

Results: Temporal zooming

Yvonne Scholz (DLR)

  • Speed-up factor: >10 reachable
  • Accuracy error of up to 35 %
slide-15
SLIDE 15

Approach II: Hardware-based computing time reduction …

By Nikitarama - Own work, CC BY-SA 4.0, https://commons.wikimedia.org/w/index.php?curid=40358482

slide-16
SLIDE 16

… and solver-based computing time reduction belong together!

Speed-Up strategies Solver-based Solver parameters Solving methodology Model-based Pure model reduction Heuristic Decomposition Exact Decomposition

slide-17
SLIDE 17

17

Annotation I

Preparing models for High Performance Computing

Reducing Model Computing Times Yvonne Scholz (DLR) August 28th, 2019

slide-18
SLIDE 18

18

Annotation II

Preparing models for High Performance Computing

Reducing Model Computing Times

  • Annotation pre-structures the optimization problem
  • The GAMS interface permutes the matrix and builds model blocks for PIPS-IPM
  • The new solver PIPS-IPM can solve the problem parallelized on a supercomputer

Yvonne Scholz (DLR) August 28th, 2019

slide-19
SLIDE 19

19

The new PIPS solver

August 28th, 2019

Results

Commercial Solvers:

Poor scaling

Time strongly depends

  • n selected solver

PIPS:

New version is much faster (note that original PIPS was developed for different problems!)

Scaling is almost linear

Still in beta state! Issues:

 parallel

preprocessing

 not suitable for all

LPs

Yvonne Scholz (DLR) Reducing Model Computing Times

slide-20
SLIDE 20

Conclusions

slide-21
SLIDE 21

21

  • Model based speed-up strategies

– Slicing / Aggregation / Heuristics / Decomposition – Computing time reduction up to factor 10

  • Solver based speed-up strategies

– ESM Annotation  GAMS interface  new PIPS solver  HPC – Computing time reduction can reach > factor 100 – New PIPS solver still in beta state

  • BEAM-ME Best Practice Guide

– publication planned by the end of 2019 – To be notified, subscribe to the mailing list: beamme-news@dlr.de subject: „subscribe“

Conclusions

Yvonne Scholz (DLR)

slide-22
SLIDE 22

Project BEAM-ME

a project by

Contact yvonne.scholz@dlr.de

Thank you!