Add support for inference with uncorrelated and correlated gaussian n…

…oise

twin4build/estimator/estimator.py

@@ -110,7 +110,7 @@ def run_emcee_estimation(self,
         assert n_cores>=1, "The argument \"n_cores\" must be larger than or equal to 1"
         assert fac_walker>=2, "The argument \"fac_walker\" must be larger than or equal to 2"
         allowed_priors = ["uniform", "gaussian"]
-        allowed_walker_initializations = ["uniform", "gaussian"]
+        allowed_walker_initializations = ["uniform", "gaussian", "ball"]
         assert prior in allowed_priors, f"The \"prior\" argument must be one of the following: {', '.join(allowed_priors)} - \"{prior}\" was provided."
         assert walker_initialization in allowed_walker_initializations, f"The \"walker_initialization\" argument must be one of the following: {', '.join(allowed_walker_initializations)} - \"{walker_initialization}\" was provided."
         assert np.all(self.x0>=self.lb), "The provided x0 must be larger than the provided lower bound lb"
@@ -137,9 +137,9 @@ def run_emcee_estimation(self,
             logprior = self.gaussian_logprior
 
         ndim = len(self.flat_attr_list)
+        self.n_par = 0
+        self.n_par_map = {}
         if assume_uncorrelated_noise==False:
-            self.n_par = 0
-            self.n_par_map = {}
             # Get number of gaussian process parameters
             for j, measuring_device in enumerate(self.targetMeasuringDevices):
                 source_component = [cp.connectsSystemThrough.connectsSystem for cp in measuring_device.connectsAt][0]
@@ -167,6 +167,12 @@ def run_emcee_estimation(self,
             ub = np.resize(self.ub,(x0_start.shape))
             x0_start[x0_start<self.lb] = lb[x0_start<self.lb]
             x0_start[x0_start>self.ub] = ub[x0_start>self.ub]
+        elif walker_initialization=="ball":
+            x0_start = np.random.uniform(low=self.x0-1e-5, high=self.ub+1e-5, size=(n_temperature, n_walkers, ndim))
+            lb = np.resize(self.lb,(x0_start.shape))
+            ub = np.resize(self.ub,(x0_start.shape))
+            x0_start[x0_start<self.lb] = lb[x0_start<self.lb]
+            x0_start[x0_start>self.ub] = ub[x0_start>self.ub]
 
         print(f"Number of cores: {n_cores}")
         print(f"Number of estimated parameters: {ndim}")

twin4build/estimator/tests/test_estimator_wbypass.py

@@ -79,9 +79,9 @@ def test_estimator():
     # Options for the PTEMCEE estimation algorithm. If the options argument is not supplied or None is supplied, default options are applied.  
     options = {"n_sample": 10000, #This is a test file, and we therefore only sample 2. Typically, we need at least 1000 samples before the chain converges. 
                 "n_temperature": 1, #Number of parallel chains/temperatures.
-                "fac_walker": 4, #Scaling factor for the number of ensemble walkers per chain. Minimum is 2.
+                "fac_walker": 8, #Scaling factor for the number of ensemble walkers per chain. Minimum is 2.
                 "prior": "uniform", #Prior distribution - "gaussian" is also implemented
-                "walker_initialization": "gaussian",#Initialization of parameters - "gaussian" is also implemented
+                "walker_initialization": "ball",#Initialization of parameters - "gaussian" is also implemented
                 "n_cores": 8,
                 "assume_uncorrelated_noise": False
                 }

twin4build/simulator/simulator.py

@@ -355,7 +355,7 @@ def _sim_func_gaussian_process(self, model, theta, stepSize, startTime, endTime,
             y_model = np.zeros((len(self.dateTimeSteps), len(self.targetMeasuringDevices)))
             y = np.zeros((len(self.dateTimeSteps), len(self.targetMeasuringDevices)))
             standardDeviation = model.chain_log["standardDeviation"]
-            n_samples = 1000
+            n_samples = 100
             n_prev = 0
             for j, measuring_device in enumerate(self.targetMeasuringDevices):
                 source_component = [cp.connectsSystemThrough.connectsSystem for cp in measuring_device.connectsAt][0]