def run(self):
"""Calculate shortest path with resource constraints.
"""
start = time()
self._init_swarm()
while (self.iter < self.max_iter
and not check_time_limit_breached(start, self.time_limit)):
pos_new = self.pos + self._get_vel()
self._update_best(self.pos, pos_new)
self.pos = pos_new
self.fitness = self._get_fitness(self.pos)
self._global_best()
# Update local best for each particle
for i in range(self.swarm_size):
self._local_best(i)
if self.iter % 100 == 0:
log.info("Iteration: {0}. Current best fit: {1}".format(
self.iter, self.best_fit))
# Terminate if feasible path found with total cost <= threshold
if (self.best_path and self.best_path_cost
and self.threshold is not None
and self.best_path_cost <= self.threshold):
break
self.iter += 1
if not self.best_path:
raise Exception("No resource feasible path has been found")
def _terminate(self) -> bool:
"""Check whether time limit is violated or final path with weight
under the input threshold"""
if self.time_limit is not None and check_time_limit_breached(
self.start_time, self.time_limit):
return True
return bool(self._check_final_label())
def run(self):
"""
Calculate shortest path with resource constraints.
"""
start = time()
while (self.it < self.max_iter and not self.stop
and not check_time_limit_breached(start, self.time_limit)):
self._algorithm()
self.it += 1
if not self.best_solution.path:
raise Exception("No resource feasible path has been found")
def run(self):
"""Calculate shortest path with resource constraints.
"""
start = time()
self._init_swarm()
while (self.iter < self.max_iter and
not check_time_limit_breached(start, self.time_limit)):
pos_new = self.pos + self._get_vel()
new_rands = self.random_state.uniform(0, 1, size=self.swarm_size)
# Force Source and Sink to be selected
pos_new[:, [0, -1]] = min(10 * self.lower_bound, np.min(pos_new))
paths_new = np.empty(len(pos_new), dtype='object')
paths_new[:] = [
self._pos2path(p, r) for p, r in zip(pos_new, new_rands)
]
self._update_best(self.pos, pos_new, self.rands, new_rands,
self.paths, paths_new)
self.pos = pos_new
self.rands = new_rands
self.fitness = self._get_fitness(self.paths)
self._global_best()
# Save best results
self.st_path = self._best_path
self.check_feasibility() # works on st_path
# Update local best for each particle
for i in range(self.swarm_size):
self._local_best(i)
if self.iter % 100 == 0:
log.info("Iteration: {0}. Current best fit: {1}".format(
self.iter, self.best_fit))
# Terminate if feasible path found with total cost <= threshold
if (self.best_path and self.best_path_cost and
self.threshold is not None and
self.best_path_cost <= self.threshold):
break
self.iter += 1
if not self.best_path:
raise Exception("No resource feasible path has been found")