def fitness(individual: IndividualType, fleet: Fleet,
hp: AlphaGA_HyperParameters):
"""
Positive fitness of the individual.
:param individual: the individual to evaluate
:param fleet: fleet instance
:param hp: GA hyper-parameters
:return: fitness value of the individual (positive)
"""
# Decode
routes = decode(individual, fleet, hp)
# Set routes
fleet.set_routes_of_vehicles(routes)
# Iterate
fleet.iterate()
# Cost
costs = np.array(fleet.cost_function()
) # cost_tt, cost_ec, cost_chg_time, cost_chg_cost
# Calculate penalization
feasible, distance, accept = fleet.feasible()
if distance > 0.:
penalization = distance + hp.hard_penalization
else:
penalization = 0
# Calculate fitness
fit = np.dot(costs, np.asarray(hp.weights)) + penalization
return fit, feasible, accept
def fitness(individual: IndividualType, indices: IndicesType, init_state: Dict,
fleet: Fleet, hp: HyperParameters):
"""
Positive fitness of the individual.
:param individual: the individual to evaluate
:param fleet: fleet instance
:param hp: GA hyper-parameters
:return: fitness value of the individual (positive)
"""
# Decode
routes = decode(individual, indices, init_state, fleet)
# Set routes
fleet.set_routes_of_vehicles(routes)
# Cost
costs = np.array(fleet.cost_function())
# Calculate penalization
feasible, distance, accept = fleet.feasible()
penalization = distance + hp.hard_penalization if fleet.deterministic else distance
# Calculate fitness
fit = np.dot(costs, np.asarray(hp.weights)) + penalization
return fit, feasible, accept
def fitness(individual: IndividualType, indices: IndicesType,
critical_points: Dict, fleet: Fleet, hp: HyperParameters):
"""
Calculates fitness of individual.
"""
# Decode
m = len(fleet.vehicles)
routes = decode(individual, indices, critical_points, fleet, hp)
# Set routes
reaching_states = {
id_ev: (r.x1_0, r.x2_0, r.x3_0)
for id_ev, r in critical_points.items()
}
vehicles_pos = [cp.S0 for cp in critical_points.values()]
init_theta = np.array(
[1 if i in vehicles_pos else 0 for i in range(len(fleet.network))])
fleet.set_routes_of_vehicles(routes,
reaching_states=reaching_states,
init_theta=init_theta)
# Cost
costs = np.array(fleet.cost_function())
# Calculate penalization
feasible, distance, accept = fleet.feasible()
penalization = 0
if not accept:
penalization = distance + hp.hard_penalization + hp.K2
elif accept and not feasible:
penalization = distance + hp.hard_penalization
# Calculate fitness
fit = np.dot(costs, np.asarray(hp.weights)) + penalization
return fit, feasible, accept