optimize_bacterialrods: main script for bacterial rods optimization

Author: polinagaindrik

examples/test_bacteriarods_optim.py

@@ -0,0 +1,52 @@
+import cr_mech_coli.crm_fit as crm_fit
+
+from cr_bayesian_optim.optimize_bacterialrods import main_bacterialrods_optimization
+from cr_bayesian_optim.plotting import *
+from cr_bayesian_optim.optimization import *
+#from skopt import gp_minimize
+
+def update_test_ABM_framework(settings):
+    n_vertices = 8
+    settings.constants.n_vertices = n_vertices
+    settings.constants.n_saves = 15
+    #settings.parameters.damping = crm_fit.SampledFloat(min=0, max=2.5, initial=1.5)
+    #settings.parameters.damping = 2.0
+    settings.parameters.potential_type.Mie.en = 10.
+    settings.parameters.potential_type.Mie.em = 1.5
+    settings.others = crm_fit.Others(True)
+    return settings
+
+'''
+def custom_optimizer(cost, bnds):
+    return gp_minimize(cost,
+                      bnds,
+                      acq_func="LCB",
+                      n_calls=20,
+                      n_random_starts=3               
+                      random_state=1234)
+    # Possible variations:
+    # - Different acquisition functions/kappa/xi/noise-?/
+    # - Different optimization methods (e.g. differential evolution)
+    # - Different parameter bounds or initial conditions / different constants in settings - ?
+'''
+
+
+if __name__ == "__main__":
+    res = main_bacterialrods_optimization(optimization_bayes)#, update_ABM=update_test_ABM_framework)
+    #res = main_bacterialrods_optimization(optimization_diff_evolution, update_ABM=update_test_ABM_framework)
+
+    # Aquisition functions: # EI/PI: faster but less accurate (?)
+                            # MES: slower, smaller cost / does not allow parallelization
+                            # PVRS: very slow
+
+    path_output =  'out/'#bacterialrods_optim/'
+    add_name = '_EI'
+
+    # Save and load optimization result
+    save_optimization_result(res, path=path_output, add_filename=add_name)
+    res = load_optimization_result(path=path_output, add_filename=add_name)
+
+    # Plotting (only for bayesian optimization):
+    plot_optimization_convergence_bayes(res, path=path_output, add_name=add_name)
+    #plot_1D_cost_approximation_bayes(res, path=path_output, add_name=add_name)
+    plot_objective_projection_bayes(res, path=path_output, add_name=add_name)