adds contour figure, shell updates, clean-up

pybop-team · Nov 26, 2024 · 1b7c54c · 1b7c54c
1 parent 183bda1
commit 1b7c54c
Show file tree

Hide file tree

Showing 30 changed files with 33 additions and 19 deletions.
diff --git a/joss/figures/contour_evolution_0.png b/joss/figures/contour_evolution_0.png
diff --git a/joss/figures/contour_evolution_1.png b/joss/figures/contour_evolution_1.png
diff --git a/joss/figures/contour_evolution_2.png b/joss/figures/contour_evolution_2.png
diff --git a/joss/figures/contour_evolution_3.png b/joss/figures/contour_evolution_3.png
diff --git a/joss/figures/contour_gradient_0.png b/joss/figures/contour_gradient_0.png
diff --git a/joss/figures/contour_gradient_1.png b/joss/figures/contour_gradient_1.png
diff --git a/joss/figures/contour_gradient_2.png b/joss/figures/contour_gradient_2.png
diff --git a/joss/figures/contour_gradient_3.png b/joss/figures/contour_gradient_3.png
diff --git a/joss/figures/contour_heuristic_0.png b/joss/figures/contour_heuristic_0.png
diff --git a/joss/figures/contour_heuristic_1.png b/joss/figures/contour_heuristic_1.png
diff --git a/joss/figures/contour_heuristic_2.png b/joss/figures/contour_heuristic_2.png
diff --git a/joss/figures/convergence_maximising.png b/joss/figures/convergence_maximising.png
diff --git a/joss/figures/convergence_minimising.png b/joss/figures/convergence_minimising.png
diff --git a/joss/figures/evolution_parameters.png b/joss/figures/evolution_parameters.png
diff --git a/joss/figures/gradient_parameters.png b/joss/figures/gradient_parameters.png
diff --git a/joss/figures/heuristic_parameters.png b/joss/figures/heuristic_parameters.png
diff --git a/joss/figures/joss/contour_total.png b/joss/figures/joss/contour_total.png
diff --git a/joss/figures/joss/converge.png b/joss/figures/joss/converge.png
diff --git a/joss/figures/joss/countour_evolution.png b/joss/figures/joss/countour_evolution.png
diff --git a/joss/figures/joss/countour_gradient.png b/joss/figures/joss/countour_gradient.png
diff --git a/joss/figures/joss/countour_heuristic.png b/joss/figures/joss/countour_heuristic.png
diff --git a/joss/figures/joss/design.png b/joss/figures/joss/design.png
diff --git a/joss/figures/joss/optimisers_parameters.png b/joss/figures/joss/optimisers_parameters.png
diff --git a/joss/figures/joss/sim-landscape.png b/joss/figures/joss/sim-landscape.png
diff --git a/joss/figures/landscape.png b/joss/figures/landscape.png
diff --git a/joss/figures/simulation.png b/joss/figures/simulation.png
diff --git a/joss/image_combine.sh b/joss/image_combine.sh
@@ -1,7 +1,15 @@
-magick convert +append figures/design_gravimetric.png figures/design_prediction.png figures/joss/design.png
+convert +append figures/design_gravimetric.png figures/design_prediction.png figures/joss/design.png
 
-magick convert +append figures/gradient_parameters.png figures/evolution_parameters.png figures/heuristic_parameters.png figures/joss/optimisers_parameters.png
+convert +append figures/gradient_parameters.png figures/evolution_parameters.png figures/heuristic_parameters.png figures/joss/optimisers_parameters.png
 
-magick convert +append figures/convergence_minimising.png figures/convergence_maximising.png figures/joss/converge.png
+convert +append figures/convergence_minimising.png figures/convergence_maximising.png figures/joss/converge.png
 
-magick convert +append figures/simulation.png figures/landscape.png figures/joss/sim-landscape.png
+convert +append figures/simulation.png figures/landscape.png figures/joss/sim-landscape.png
+
+convert +append figures/contour_gradient_0.png figures/contour_gradient_1.png figures/contour_gradient_2.png figures/contour_gradient_3.png figures/joss/countour_gradient.png
+
+convert +append figures/contour_evolution_0.png figures/contour_evolution_1.png figures/contour_evolution_2.png figures/contour_evolution_3.png figures/joss/countour_evolution.png
+
+convert +append figures/contour_heuristic_0.png figures/contour_heuristic_1.png figures/contour_heuristic_2.png figures/joss/countour_heuristic.png
+
+convert -append figures/joss/countour_gradient.png figures/joss/countour_evolution.png figures/joss/countour_heuristic.png figures/joss/countour_total.png
diff --git a/joss/paper.md b/joss/paper.md
@@ -152,6 +152,8 @@ Furthermore, we can also compare the performance of the various optimisation alg
 
 ![Convergence in the parameter values obtained for various (meta)heuristics.  \label{fig:optimiser-inference}](figures/joss/optimisers_parameters.png){ width=100% }
 
+![Counter plot for all optimisers.  \label{fig:optimiser-inference}](figures/joss/contour_total.png){ width=100% }
+
 This parameterisation task can also be approached from the Bayesian perspective, which we will present below using PyBOP's sampler methods. The optimisation equation presented in equation \autoref{eqn:parameterisation} does not represent the Bayesian parameter identification task, and as such we introduce Bayes Theorem as,
 
 \begin{equation}

diff --git a/joss/paper.pdf b/joss/paper.pdf
diff --git a/joss/param_plots.py b/joss/param_plots.py
@@ -185,7 +185,6 @@
         # Run optimisation
         results = optim.run()
         print("True parameter values:", parameters.true_value())
-        print("Estimated parameters:", results.x)
 
         # Plot convergence
         cost_log = optim.log["cost_best"]
@@ -220,13 +219,7 @@
     ## Do the same for the maximising cost functions
     convergence_traces = []
     first_MAP = True
-    # sigma0 = pybop.Parameter(
-    #     "Sigma",
-    #     initial_value=0.012,
-    #     prior=pybop.Gaussian(0.012, 0.005),
-    #     bounds=[0, 0.05],
-    # )
-    # sigma0 = 0.02
+
     for cost in maximising_cost_classes:
         if cost is pybop.GaussianLogLikelihoodKnownSigma:
             cost = cost(problem, sigma0=sigma)
@@ -246,14 +239,13 @@
         optim = pybop.CMAES(
             cost,
             verbose=True,
-            max_iterations=500,
+            max_iterations=125,
             max_unchanged_iterations=25,
         )
 
         # Run optimisation
         results = optim.run()
         print("True parameter values:", parameters.true_value())
-        print("Estimated parameters:", results.x)
 
         # Plot convergence
         cost_log = optim.log["cost_best"]
@@ -342,7 +334,6 @@
         # Run optimisation
         results = optim.run()
         print("True parameter values:", parameters.true_value())
-        print("Estimated parameters:", results.x)
 
         # Plot the parameter traces
         parameter_fig = pybop.plot.parameters(optim, show=False, title=None)
@@ -355,6 +346,12 @@
                 parameter_fig.data[j].showlegend = False
         parameter_traces.extend(parameter_fig.data)
 
+        # Plot the cost landscape with optimisation path
+        contour = pybop.plot.contour(
+            optim, steps=25, title="", margin=dict(l=30, r=30, t=30, b=30)
+        )
+        contour.write_image(f"joss/figures/contour_gradient_{i}.png")
+
     # Plot the parameter traces together
     parameter_fig.update_layout(width=576, height=1024, plot_bgcolor="white")
     parameter_fig.data = []
@@ -396,7 +393,6 @@
         # Run optimisation
         results = optim.run()
         print("True parameter values:", parameters.true_value())
-        print("Estimated parameters:", results.x)
 
         # Plot the parameter traces
         parameter_fig = pybop.plot.parameters(optim, show=False, title=None)
@@ -409,6 +405,12 @@
                 parameter_fig.data[j].showlegend = False
         parameter_traces.extend(parameter_fig.data)
 
+        # Plot the cost landscape with optimisation path
+        contour = pybop.plot.contour(
+            optim, steps=25, title="", margin=dict(l=30, r=30, t=30, b=30)
+        )
+        contour.write_image(f"joss/figures/contour_evolution_{i}.png")
+
     # Plot the parameter traces together
     parameter_fig.update_layout(width=576, height=1024, plot_bgcolor="white")
     parameter_fig.data = []
@@ -440,7 +442,6 @@
         # Run optimisation
         results = optim.run()
         print("True parameter values:", parameters.true_value())
-        print("Estimated parameters:", results.x)
 
         # Plot the parameter traces
         parameter_fig = pybop.plot.parameters(optim, show=False, title=None)
@@ -453,8 +454,11 @@
                 parameter_fig.data[j].showlegend = False
         parameter_traces.extend(parameter_fig.data)
 
-        # # Plot the cost landscape with optimisation path
-        # pybop.plot.contour(optim, steps=15)
+        # Plot the cost landscape with optimisation path
+        contour = pybop.plot.contour(
+            optim, steps=25, title="", margin=dict(l=30, r=30, t=30, b=30)
+        )
+        contour.write_image(f"joss/figures/contour_heuristic_{i}.png")
 
     # Plot the parameter traces together
     parameter_fig.update_layout(width=576, height=1024, plot_bgcolor="white")