def __getCandidates__(self, **kwargs) -> list:
"""
Description:
Returns a list of candidates for the specified algorithm.
|\n
|\n
|\n
|\n
|\n
Args:
+ optimizer = ( enum ) The optimzier
~ Required
+ params = ( dict ) The optimizer's parameters
~ Required
+ initial = ( list ) The initial candidate
+ region = ( float ) The algorithm's current region
~ Required if <optimizer="tro">
|\n
Returns
+ list = ( list ) A list of new candidates
"""
# STEP 0: Local variables
lCandidates = []
# STEP 1: Setup - Local variables
# STEP 2: Check if optimizer passed
if ("optimizer" not in kwargs):
# STEP 3: Error handling
raise Exception(
"An error occured in SpongeBob.__getCandidates__() -> Step 2: No optimizer passed"
)
# STEP 4: Check optimizer parameters passed
if ("params" not in kwargs):
# STEP 5: Error handling
raise Exception(
"An error occured in SpongeBob.__getCandidates__() -> Step 4: No optimizer parameters passed"
)
# STEP 6: Check if initial candidate passed
if ("initial" not in kwargs):
# STEP 7: Error handling
raise Exception(
"An error occured in SpongeBob.__getCandidates__() -> Step 6: No initial candidate passed"
)
# STEP 8: Check if optimizer is tro
if (kwargs["optimizer"] == ga.TRO):
# STEP 9: Check if region passed
if ("region" not in kwargs):
# STEP 10: Error handling
raise Exception(
"An error occured in SpongeBob.__getCandidates__() -> Step 9: No region passed"
)
# STEP 11: Iterate through the required number of candidates
for _ in range(0, kwargs["params"]["candidates"]):
# STEP 12: Get temporary candidate
lTmp = Helga.getShape(kwargs["initial"])
# STEP 13: Iterate through candidate
for i in range(0, len(lTmp)):
# STEP 14: Check if single point
if (type(lTmp[i]) == float):
# STEP 16: Modify value using region and scalar
lTmp[i] = rn.gauss(
kwargs["initial"][i],
kwargs["params"]["scalar"] * kwargs["region"])
# STEP 17: Check if value above upper limit
if (lTmp[i] > 1.0):
# STEP 18: Limit value
lTmp[i] = 1.0
# STEP 19: Check if value below lower limit
if (lTmp[i] < -1.0):
# STEP 20: Limit value
lTmp[i] = -1.0
else:
# STEP 21: Iterate through list
for j in range(0, len(lTmp[i])):
# STEP 22: Modify value using region and sacalar
lTmp[i][j] = rn.gauss(
kwargs["initial"][i][j],
kwargs["params"]["scalar"] * kwargs["region"])
# STEP 23: Check if value above upper limit
if (lTmp[i][j] > 1.0):
# STEP 24: Limit value
lTmp[i][j] = 1.0
# STEP 25: Check if value below lower limit
if (lTmp[i][j] < -1.0):
# STEP 26: Limit value
lTmp[i][j] = -1.0
# STEP 22: Append new candidate to output list
lCandidates.append(lTmp)
# STEP 23: Return
return lCandidates
# STEP ??: Error handling
raise Exception(
"An error occured in SpongeBob.__getCandidates__(): Unimplemented optimizer passed"
)
def __getCandidates__(self, **kwargs) -> list:
"""
Description:
Returns a list of candidates for the specified algorithm.
|\n
|\n
|\n
|\n
|\n
Args:
+ optimizer = ( enum ) The optimizer
~ Required
+ params = ( dict ) The optimizer's parameters
~ Required
+ initial = ( list ) The initial candidate
~ Required
|\n
Returns:
+ list = ( list ) A list of new candidates
"""
# STEP 0: Local variables
lCandidates = []
lShape = None
# STEP 1: Setup - Local variables
lShape = Helga.getShape(kwargs["initial"])
# region STEP 2->??: PSO Candidate generation
# STEP 2: Check if PSO
if (kwargs["optimizer"] == sw.PSO):
# STEP 3: Append the initial candidate to the candidate list
lCandidates.append(kwargs["initial"])
# STEP 4: Iterate through remaining required candidates
for _ in range(1, kwargs["params"]["candidates"]):
# STEP 5: Get new empty candidate
lTmp = cp.deepcopy(lShape)
# STEP 6: Iterate through candidate list
for i in range(0, len(lTmp)):
# STEP 7: Check if single data point
if (type(lTmp[i]) == float):
# STEP 8: Set value
lTmp[i] = rn.uniform(-1.0, 1.0)
else:
# STEP 9: Iterate through list
for j in range(0, len(lTmp[i])):
# STEP 10: Set value
lTmp[i][j] = rn.uniform(-1.0, 1.0)
# STEP 11: Append the candidate to the output list
lCandidates.append(lTmp)
# STEP 12: Return
return lCandidates
# STEP ??: Error handling
raise Exception("An error occured in Sarah.__getCandidates__(): Unimplemented optimizer passed")
