Updated thesis

Reports/Thesis/sections/appendix.aux

@@ -1,12 +1,12 @@
 \relax 
 \providecommand\hyper@newdestlabel[2]{}
-\@writefile{toc}{\contentsline {section}{\numberline {A}Appendix}{57}{appendix.A}\protected@file@percent }
+\@writefile{toc}{\contentsline {section}{\numberline {A}Appendix}{56}{appendix.A}\protected@file@percent }
-\@writefile{lof}{\contentsline {figure}{\numberline {21}{\ignorespaces Comparison of the autoregressive models with the diffusion model\relax }}{57}{figure.caption.35}\protected@file@percent }
+\@writefile{lof}{\contentsline {figure}{\numberline {21}{\ignorespaces Comparison of the autoregressive models with the diffusion model\relax }}{56}{figure.caption.35}\protected@file@percent }
-\newlabel{fig:ar_linear_gru_comparison}{{21}{57}{Comparison of the autoregressive models with the diffusion model\relax }{figure.caption.35}{}}
+\newlabel{fig:ar_linear_gru_comparison}{{21}{56}{Comparison of the autoregressive models with the diffusion model\relax }{figure.caption.35}{}}
-\@writefile{lof}{\contentsline {figure}{\numberline {22}{\ignorespaces Comparison of the non-autoregressive models with the diffusion model\relax }}{58}{figure.caption.36}\protected@file@percent }
+\@writefile{lof}{\contentsline {figure}{\numberline {22}{\ignorespaces Comparison of the non-autoregressive models with the diffusion model\relax }}{57}{figure.caption.36}\protected@file@percent }
-\newlabel{fig:ar_linear_gru_comparison}{{22}{58}{Comparison of the non-autoregressive models with the diffusion model\relax }{figure.caption.36}{}}
+\newlabel{fig:ar_linear_gru_comparison}{{22}{57}{Comparison of the non-autoregressive models with the diffusion model\relax }{figure.caption.36}{}}
 \@setckpt{sections/appendix}{
-\setcounter{page}{59}
+\setcounter{page}{58}
 \setcounter{equation}{8}
 \setcounter{enumi}{0}
 \setcounter{enumii}{0}

Reports/Thesis/sections/results/policies/nrv_samples_policy.tex

@@ -130,11 +130,9 @@ A comparison of the baselines and the best-performing models is shown in Table \
         \bottomrule
     \end{tabular}
     \end{adjustbox}
-    \caption{Comparison of the different models using the CRPS, profit, charge cycles and penalty. The best-performing models for a certain type are selected based on the profit.}
+    \caption{Comparison of the profit achieved by the baselines and the best-performing models. The improvement is calculated compared to the baseline that uses the NRV of yesterday as a prediction.}
     \label{tab:policy_comparison}
 \end{table}
 
-\newpage
-
 \section{Conclusion}
 In this thesis, generative methods are explored to model the NRV data of the Belgian electricity market. These methods are then used to improve the decision-making to charge and discharge a battery to make a profit.  

Reports/Thesis/verslag.aux

@@ -113,9 +113,9 @@
 \newlabel{tab:diffusion_policy_comparison}{{13}{49}{Comparison of diffusion models using different hyperparameters. Early stopping is done based on the profit using the validation set.\relax }{table.caption.32}{}}
 \@writefile{lof}{\contentsline {figure}{\numberline {20}{\ignorespaces Comparison of the two samples from the model with the lowest CRPS and the model with the highest profit. \relax }}{50}{figure.caption.33}\protected@file@percent }
 \newlabel{fig:diffusion_policy_comparison_high_low_crps}{{20}{50}{Comparison of the two samples from the model with the lowest CRPS and the model with the highest profit. \relax }{figure.caption.33}{}}
-\@writefile{lot}{\contentsline {table}{\numberline {14}{\ignorespaces Comparison of the different models using the CRPS, profit, charge cycles and penalty. The best-performing models for a certain type are selected based on the profit.\relax }}{51}{table.caption.34}\protected@file@percent }
+\@writefile{lot}{\contentsline {table}{\numberline {14}{\ignorespaces Comparison of the profit achieved by the baselines and the best-performing models. The improvement is calculated compared to the baseline that uses the NRV of yesterday as a prediction.\relax }}{51}{table.caption.34}\protected@file@percent }
-\newlabel{tab:policy_comparison}{{14}{51}{Comparison of the different models using the CRPS, profit, charge cycles and penalty. The best-performing models for a certain type are selected based on the profit.\relax }{table.caption.34}{}}
+\newlabel{tab:policy_comparison}{{14}{51}{Comparison of the profit achieved by the baselines and the best-performing models. The improvement is calculated compared to the baseline that uses the NRV of yesterday as a prediction.\relax }{table.caption.34}{}}
-\@writefile{toc}{\contentsline {section}{\numberline {7}Conclusion}{52}{section.7}\protected@file@percent }
+\@writefile{toc}{\contentsline {section}{\numberline {7}Conclusion}{51}{section.7}\protected@file@percent }
 \bibstyle{unsrtnat}
 \bibdata{references}
 \bibcite{commission_for_electricity_and_gas_regulation_creg_study_2023}{{1}{}{{Commission for Electricity and Gas Regulation (CREG)}}{{}}}
@@ -189,4 +189,4 @@
 \ACRO{pages}{MSE=={44@1@43|45@1@44}}
 \ACRO{pages}{MAE=={44@1@43|45@1@44}}
 \ACRO{pages}{NRV=={3@1@2|4@1@3|47@1@46|48@1@47}}
-\gdef \@abspage@last{59}
+\gdef \@abspage@last{58}

Reports/Thesis/verslag.log

@@ -1,4 +1,4 @@
-This is pdfTeX, Version 3.141592653-2.6-1.40.25 (TeX Live 2023) (preloaded format=pdflatex 2023.9.17)  19 MAY 2024 23:59
+This is pdfTeX, Version 3.141592653-2.6-1.40.25 (TeX Live 2023) (preloaded format=pdflatex 2023.9.17)  20 MAY 2024 00:35
 entering extended mode
  restricted \write18 enabled.
  file:line:error style messages enabled.
@@ -1571,7 +1571,7 @@ File: images/diffusion/policies/comparison/Testing_7008_High_CRPS.jpeg Graphic f
 <use images/diffusion/policies/comparison/Testing_7008_High_CRPS.jpeg>
 Package pdftex.def Info: images/diffusion/policies/comparison/Testing_7008_High_CRPS.jpeg  used on input line 99.
 (pdftex.def)             Requested size: 223.07211pt x 111.11894pt.
- [49] [50 <./images/diffusion/policies/comparison/Testing_864_Low_CRPS.jpeg> <./images/diffusion/policies/comparison/Testing_864_High_CRPS.jpeg> <./images/diffusion/policies/comparison/Testing_7008_Low_CRPS.jpeg> <./images/diffusion/policies/comparison/Testing_7008_High_CRPS.jpeg>] [51])) [52] [53] (./verslag.bbl
+ [49] [50 <./images/diffusion/policies/comparison/Testing_864_Low_CRPS.jpeg> <./images/diffusion/policies/comparison/Testing_864_High_CRPS.jpeg> <./images/diffusion/policies/comparison/Testing_7008_Low_CRPS.jpeg> <./images/diffusion/policies/comparison/Testing_7008_High_CRPS.jpeg>])) [51] [52] (./verslag.bbl
 LaTeX Font Info:    Trying to load font information for T1+LinuxLibertineMonoT-TLF on input line 11.
  (/usr/local/texlive/2023/texmf-dist/tex/latex/libertine/T1LinuxLibertineMonoT-TLF.fd
 File: T1LinuxLibertineMonoT-TLF.fd 2019/05/20 (autoinst) Font definitions for T1/LinuxLibertineMonoT-TLF.
@@ -1640,7 +1640,7 @@ Underfull \hbox (badness 1635) in paragraph at lines 49--53
 |   []   |   []   |   []\T1/LinuxLibertineT-TLF/m/n/12 , vol-ume 37 of |   []   |   []   |   []   |   []   |   [],
  []
 
-[54{/usr/local/texlive/2023/texmf-dist/fonts/enc/dvips/libertine/lbtn_dhi6qf.enc}]
+[53{/usr/local/texlive/2023/texmf-dist/fonts/enc/dvips/libertine/lbtn_dhi6qf.enc}]
 Underfull \hbox (badness 1430) in paragraph at lines 74--80
 \T1/LinuxLibertineT-TLF/m/n/12 and es-ti-ma-tion.  102(477):359--378.  ISSN 0162-1459, 1537-274X.  doi: $10 . 1198 /
  []
@@ -1650,7 +1650,7 @@ Underfull \hbox (badness 2134) in paragraph at lines 123--129
 \T1/LinuxLibertineT-TLF/m/n/12 117871$.  URL [][]$\T1/LinuxLibertineMonoT-TLF/regular/n/12 https : / / linkinghub . elsevier . com / retrieve / pii /
  []
 
-[55]
+[54]
 Underfull \hbox (badness 10000) in paragraph at lines 152--154
 \T1/LinuxLibertineMonoT-TLF/regular/n/12 elia . be / explore / dataset / ods047 / information / ?sort =
  []
@@ -1685,7 +1685,7 @@ Underfull \hbox (badness 10000) in paragraph at lines 168--170
 \T1/LinuxLibertineMonoT-TLF/regular/n/12 elia . be / explore / dataset / ods022 / information / ?sort =
  []
 
-) [56]
+) [55]
 \openout2 = `sections/appendix.aux'.
 
  (./sections/appendix.tex
@@ -1817,10 +1817,10 @@ File: images/diffusion/results/samples/Diffusion_Test_Example_7008.jpeg Graphic
 <use images/diffusion/results/samples/Diffusion_Test_Example_7008.jpeg>
 Package pdftex.def Info: images/diffusion/results/samples/Diffusion_Test_Example_7008.jpeg  used on input line 154.
 (pdftex.def)             Requested size: 172.99498pt x 86.16971pt.
- [57
+ [56
 
 
-] [58]) (./verslag.aux (./sections/introduction.aux) (./sections/background.aux) (./sections/policies.aux) (./sections/literature_study.aux) (./sections/appendix.aux))
+] [57]) (./verslag.aux (./sections/introduction.aux) (./sections/background.aux) (./sections/policies.aux) (./sections/literature_study.aux) (./sections/appendix.aux))
 
 LaTeX Warning: There were undefined references.
 
@@ -1831,18 +1831,18 @@ Package rerunfilecheck Info: File `verslag.out' has not changed.
 (rerunfilecheck)             Checksum: 78A3B45A8B1D632CBDB591FF4FE72E7D;5101.
  ) 
 Here is how much of TeX's memory you used:
- 35786 strings out of 476025
+ 35785 strings out of 476025
- 784747 string characters out of 5790017
+ 784740 string characters out of 5790017
  1883388 words of memory out of 5000000
  55362 multiletter control sequences out of 15000+600000
  613012 words of font info for 120 fonts, out of 8000000 for 9000
  1141 hyphenation exceptions out of 8191
  84i,16n,131p,1821b,765s stack positions out of 10000i,1000n,20000p,200000b,200000s
  </Users/victormylle/Library/texlive/2023/texmf-var/fonts/pk/ljfour/public/bbm/bbm12.600pk></usr/local/texlive/2023/texmf-dist/fonts/type1/public/libertinust1math/LibertinusT1Math.pfb></usr/local/texlive/2023/texmf-dist/fonts/type1/public/libertine/LinBiolinumT.pfb></usr/local/texlive/2023/texmf-dist/fonts/type1/public/libertine/LinBiolinumTB.pfb></usr/local/texlive/2023/texmf-dist/fonts/type1/public/libertine/LinLibertineMT.pfb></usr/local/texlive/2023/texmf-dist/fonts/type1/public/libertine/LinLibertineT.pfb></usr/local/texlive/2023/texmf-dist/fonts/type1/public/libertine/LinLibertineTB.pfb></usr/local/texlive/2023/texmf-dist/fonts/type1/public/libertine/LinLibertineTI.pfb></usr/local/texlive/2023/texmf-dist/fonts/type1/public/stix/stix-mathcal.pfb>
-Output written on verslag.pdf (59 pages, 9101254 bytes).
+Output written on verslag.pdf (58 pages, 9101037 bytes).
 PDF statistics:
- 857 PDF objects out of 1000 (max. 8388607)
+ 853 PDF objects out of 1000 (max. 8388607)
- 662 compressed objects within 7 object streams
+ 659 compressed objects within 7 object streams
- 168 named destinations out of 1000 (max. 500000)
+ 167 named destinations out of 1000 (max. 500000)
  605 words of extra memory for PDF output out of 10000 (max. 10000000)
 

Reports/Thesis/verslag.toc

@@ -32,5 +32,5 @@
 \contentsline {subsection}{\numberline {6.5}Policies for battery optimization}{46}{subsection.6.5}%
 \contentsline {subsubsection}{\numberline {6.5.1}Baselines}{46}{subsubsection.6.5.1}%
 \contentsline {subsubsection}{\numberline {6.5.2}Policy using generated NRV samples}{47}{subsubsection.6.5.2}%
-\contentsline {section}{\numberline {7}Conclusion}{52}{section.7}%
+\contentsline {section}{\numberline {7}Conclusion}{51}{section.7}%
-\contentsline {section}{\numberline {A}Appendix}{57}{appendix.A}%
+\contentsline {section}{\numberline {A}Appendix}{56}{appendix.A}%

src/trainers/quantile_trainer.py

@@ -637,6 +637,15 @@ class NonAutoRegressiveQuantileRegression(Trainer):
                 task.get_logger().report_single_value(
                     name=metric_name, value=metric_value
                 )
+            
+            if train == False:
+                crps_from_samples_metric, self.test_set_samples = (
+                    self.calculate_crps_from_samples(None, dataloader, None)
+                )
+                task.get_logger().report_single_value(
+                    name="test_CRPS_from_samples_transformed",
+                    value=np.mean(crps_from_samples_metric),
+                )
 
     def debug_plots(self, task, train: bool, data_loader, sample_indices, epoch):
         for actual_idx, idx in sample_indices.items():
@@ -846,6 +855,7 @@ class NonAutoRegressiveQuantileRegression(Trainer):
                     targets = targets.squeeze(-1)
                     targets = targets[0].unsqueeze(0)
                     targets = self.data_processor.inverse_transform(targets)
+                    targets = targets.to(self.device)
 
                     samples = samples.to(self.device)
 
@@ -871,8 +881,8 @@ class NonAutoRegressiveQuantileRegression(Trainer):
                         initial_penalty=900,
                         target_charge_cycles=58 * 400 / 356,
                         initial_learning_rate=5,
-                        max_iterations=100,
+                        max_iterations=30,
-                        tolerance=1,
+                        tolerance=2,
                         iteration=epoch,
                     )
                 )
@@ -907,6 +917,8 @@ class NonAutoRegressiveQuantileRegression(Trainer):
                     generated_samples,
                 )
 
+        return np.mean(crps_from_samples_metric), generated_samples
+
     def plot_quantile_percentages(
         self,
         task,

src/training_scripts/autoregressive_quantiles.py

@@ -44,7 +44,7 @@ data_config.NOMINAL_NET_POSITION = True
 
 data_config = task.connect(data_config, name="data_features")
 
-data_processor = DataProcessor(data_config, path="", lstm=True)
+data_processor = DataProcessor(data_config, path="", lstm=False)
 data_processor.set_batch_size(512)
 data_processor.set_full_day_skip(False)
 

src/training_scripts/non_autoregressive_quantiles.py

@@ -2,10 +2,10 @@ from src.utils.clearml import ClearMLHelper
 
 #### ClearML ####
 clearml_helper = ClearMLHelper(
-    project_name="Thesis/NAQR: Non Linear (4 - 256) + Load + PV + Wind + NP"
+    project_name="Thesis/NrvForecast"
 )
 task = clearml_helper.get_task(
-    task_name="NAQR: Non Linear (4 - 256) + Load + PV + Wind + NP"
+    task_name="NAQR: Non Linear (2 - 512)"
 )
 task.execute_remotely(queue_name="default", exit_process=True)
 
@@ -30,17 +30,17 @@ from src.models.time_embedding_layer import TimeEmbedding
 #### Data Processor ####
 data_config = DataConfig()
 
-data_config.NRV_HISTORY = True
+data_config.NRV_HISTORY = False
-data_config.LOAD_HISTORY = True
+data_config.LOAD_HISTORY = False
-data_config.LOAD_FORECAST = True
+data_config.LOAD_FORECAST = False
 
-data_config.WIND_FORECAST = True
+data_config.WIND_FORECAST = False
-data_config.WIND_HISTORY = True
+data_config.WIND_HISTORY = False
 
-data_config.PV_FORECAST = True
+data_config.PV_FORECAST = False
-data_config.PV_HISTORY = True
+data_config.PV_HISTORY = False
 
-data_config.NOMINAL_NET_POSITION = True
+data_config.NOMINAL_NET_POSITION = False
 
 
 data_config = task.connect(data_config, name="data_features")
@@ -53,7 +53,7 @@ data_processor.set_full_day_skip(True)
 #### Hyperparameters ####
 data_processor.set_output_size(96)
 inputDim = data_processor.get_input_size()
-epochs = 300
+epochs = 5
 
 # add parameters to clearml
 quantiles = task.get_parameter("general/quantiles", cast=True)
@@ -69,7 +69,7 @@ else:
 model_parameters = {
     "learning_rate": 0.0001,
     "hidden_size": 512,
-    "num_layers": 8,
+    "num_layers": 2,
     "dropout": 0.2,
 }
 
@@ -111,15 +111,15 @@ trainer = NonAutoRegressiveQuantileRegression(
     data_processor,
     quantiles,
     "cuda",
-    policy_evaluator=None,
+    policy_evaluator=policy_evaluator,
     debug=False,
 )
 
 trainer.add_metrics_to_track(
     [PinballLoss(quantiles), MSELoss(), L1Loss(), CRPSLoss(quantiles)]
 )
-trainer.early_stopping(patience=5)
+trainer.early_stopping(patience=8)
-trainer.plot_every(20)
+trainer.plot_every(4)
 trainer.train(task=task, epochs=epochs, remotely=True)
 
 ### Policy Evaluation ###
@@ -138,7 +138,7 @@ optimal_penalty, profit, charge_cycles = (
         test_loader=test_loader,
         initial_penalty=1000,
         target_charge_cycles=283,
-        learning_rate=15,
+        initial_learning_rate=15,
         max_iterations=150,
         tolerance=1,
     )