Update main.py

main.py

@@ -13,21 +13,12 @@ import numpy_financial as npf
 
 optuna.logging.set_verbosity(optuna.logging.WARNING)  
 
-FAST_MODE = True
 PROJECT_YEARS = 15
 BASE_CAPACITY_KTA = 300
 INFLATION_RATE = 0.015
 TAX_RATE = 0.10
 DEPRECIATION_YEARS = 15
 
-
-n_calls = 45 if FAST_MODE else 50
-n_trials = 90 if FAST_MODE else 100
-population_size = 40 if FAST_MODE else 50
-generations = 80 if FAST_MODE else 100
-
-
-
 TECHNOLOGY_DATA = {
     "JNC": {"capex_base_M": 180.0, "opex_base_cents_kg": 150.0},
     "Hoechst_AG": {"capex_base_M": 230.0, "opex_base_cents_kg": 155.0},
@@ -59,7 +50,7 @@ PRODUCT_PRICES_USD_PER_TON = {
 
 OPTIMIZATION_SPACE = {
     'capacity_kta': (500, 600),
-    'technology': ["Engro_Pakistan", "Shin_Etsu_2004"],
+    'technology': ["Engro_Pakistan"], #, "Shin_Etsu_2004"],
     'sourcing_strategy': ['Integrated_Production'],
     'export_market_mix': (0.6, 0.8),
     'sell_byproducts': [True]
@@ -129,14 +120,12 @@ def run_optimizations_without_ml():
     results.append(run_optuna_direct())
     return results
 
-
-
 def run_genetic_algorithm():
     print("Running Genetic Algorithm (GA)...")
-    population = [{k: random.choice(v) if isinstance(v, list) else random.uniform(*v) for k,v in OPTIMIZATION_SPACE.items()} for _ in range(population_size)]
+    population = [{k: random.choice(v) if isinstance(v, list) else random.uniform(*v) for k,v in OPTIMIZATION_SPACE.items()} for _ in range(50)]
     best_overall_individual = None
     best_overall_fitness = -float('inf')
-    for _ in range(generations):
+    for _ in range(100):
         fitnesses = [calculate_project_kpis(**ind)['irr'] for ind in population]
         if max(fitnesses) > best_overall_fitness:
             best_overall_fitness = max(fitnesses)
@@ -165,7 +154,7 @@ def run_bayesian_optimization():
     @use_named_args(skopt_space)
     def objective(**params):
         return -calculate_project_kpis(**params)['irr']
-    res = gp_minimize(objective, skopt_space, n_calls=n_calls, random_state=42, n_initial_points=20)
+    res = gp_minimize(objective, skopt_space, n_calls=100, random_state=42, n_initial_points=20)
     best_params = {space.name: val for space, val in zip(skopt_space, res.x)}
     kpis = calculate_project_kpis(**best_params)
     return {"Method": "Bayesian Opt", **kpis, "Params": best_params}
@@ -182,7 +171,7 @@ def run_optuna_direct():
         }
         return calculate_project_kpis(**params)['irr']
     study = optuna.create_study(direction="maximize")
-    study.optimize(objective, n_trials=n_trials, n_jobs=-1)
+    study.optimize(objective, n_trials=200, n_jobs=-1)
     kpis = calculate_project_kpis(**study.best_params)
     return {"Method": "Optuna (TPE - Direct)", **kpis, "Params": study.best_params}