def anytime_gbfs(initial_state, heur_fn, timebound=10):
# IMPLEMENT
"""Provides an implementation of anytime greedy best-first search, as described in the HW1 handout"""
"""INPUT: a rush hour state that represents the start state and a timebound (number of seconds)"""
"""OUTPUT: A goal state (if a goal is found), else False"""
"""implementation of anytime greedybfs algorithm"""
search_engine = SearchEngine("best_first", "full")
search_engine.init_search(initial_state, rushhour_goal_fn, heur_fn)
gval_cost_bound = float("inf")
time_left = timebound
init_time = os.times()[0]
solution = search_engine.search(
timebound=time_left, costbound=(gval_cost_bound, float("inf"), float("inf"))
)
finish_time = os.times()[0]
time_left -= finish_time - init_time
if solution:
gval_cost_bound = solution.gval
else:
return False
while time_left > 0:
init_time = os.times()[0]
improved_solution = search_engine.search(
timebound=time_left, costbound=(gval_cost_bound, float("inf"), float("inf"))
)
time_left -= os.times()[0] - init_time
if improved_solution:
gval_cost_bound = improved_solution.gval
solution = improved_solution
else:
break
return solution
def anytime_weighted_astar(initial_state, heur_fn, weight=1.0, timebound=10):
# IMPLEMENT
"""Provides an implementation of anytime weighted a-star, as described in the HW1 handout"""
"""INPUT: a rush hour state that represents the start state and a timebound (number of seconds)"""
"""OUTPUT: A goal state (if a goal is found), else False"""
"""implementation of weighted astar algorithm"""
time_left = timebound
wrapped_fval_function = lambda sN: fval_function(sN, weight)
se = SearchEngine("custom", "full")
se.init_search(initial_state, rushhour_goal_fn, heur_fn, wrapped_fval_function)
cost_bound = float("inf")
init_time = os.times()[0]
solution = se.search(
timebound=time_left, costbound=(float("inf"), float("inf"), cost_bound)
)
finish_time = os.times()[0]
time_left -= finish_time - init_time
if solution:
cost_bound = solution.gval + heur_fn(solution)
else:
return False
while time_left > 0:
init_time = os.times()[0]
improved_solution = se.search(
timebound=time_left, costbound=(float("inf"), float("inf"), cost_bound)
)
finish_time = os.times()[0]
time_left -= finish_time - init_time
if improved_solution:
cost_bound = improved_solution.gval + heur_fn(improved_solution)
solution = improved_solution
else:
break
return solution