def __calculate_heuristics(prev_h, prev_x, m0, index, corresp, t0, sync_net, incidence_matrix, fin_vec, cost_vec, a_matrix, g_matrix, h_cvx, variant, use_cvxopt=False, compute_exact_heu=False): """ Calculate the heuristics Returns --------------- h Heuristic value x Solution """ m, t = get_corresp_marking_and_trans(m0, index, corresp, t0) h = 0 x = None if compute_exact_heu or t is None: h, x = utils.__compute_exact_heuristic_new_version(sync_net, a_matrix, h_cvx, g_matrix, cost_vec, incidence_matrix, m, fin_vec, variant, use_cvxopt=use_cvxopt) trustable = True else: h, x = utils.__derive_heuristic(incidence_matrix, cost_vec, prev_x, t, prev_h) trustable = utils.__trust_solution(x) return h, x, trustable
def __search(sync_net, ini, fin, stop, cost_function, skip, max_trace_length): from pm4py.objects.petri.utils import decorate_places_preset_trans, decorate_transitions_prepostset decorate_transitions_prepostset(sync_net) decorate_places_preset_trans(sync_net) incidence_matrix = petri.incidence_matrix.construct(sync_net) ini_vec, fin_vec, cost_vec = utils.__vectorize_initial_final_cost( incidence_matrix, ini, fin, cost_function) closed = set() a_matrix = np.asmatrix(incidence_matrix.a_matrix).astype(np.float64) g_matrix = -np.eye(len(sync_net.transitions)) h_cvx = np.matrix(np.zeros(len(sync_net.transitions))).transpose() cost_vec = [x * 1.0 for x in cost_vec] use_cvxopt = False if lp_solver_factory.DEFAULT_LP_SOLVER_VARIANT == lp_solver_factory.CVXOPT_SOLVER_CUSTOM_ALIGN or lp_solver_factory.DEFAULT_LP_SOLVER_VARIANT == lp_solver_factory.CVXOPT_SOLVER_CUSTOM_ALIGN_ILP: use_cvxopt = True if use_cvxopt: # not available in the latest version of PM4Py from cvxopt import matrix a_matrix = matrix(a_matrix) g_matrix = matrix(g_matrix) h_cvx = matrix(h_cvx) cost_vec = matrix(cost_vec) h, x = utils.__compute_exact_heuristic_new_version( sync_net, a_matrix, h_cvx, g_matrix, cost_vec, incidence_matrix, ini, fin_vec, lp_solver_factory.DEFAULT_LP_SOLVER_VARIANT, use_cvxopt=use_cvxopt) # heuristics need to be adapted for prefix alignments # here we make the heuristics way less powerful h = h / (max_trace_length + 1.0) ini_state = SearchTuple(0 + h, 0, h, ini, None, None, x, True) open_set = [ini_state] heapq.heapify(open_set) visited = 0 queued = 0 traversed = 0 while not len(open_set) == 0: curr = heapq.heappop(open_set) current_marking = curr.m # 11/10/2019 (optimization Y, that was optimization X, # but with the good reasons this way): avoid checking markings in the cycle using # the __get_alt function, but check them 'on the road' already_closed = current_marking in closed if already_closed: continue while not curr.trust: h, x = utils.__compute_exact_heuristic_new_version( sync_net, a_matrix, h_cvx, g_matrix, cost_vec, incidence_matrix, curr.m, fin_vec, lp_solver_factory.DEFAULT_LP_SOLVER_VARIANT, use_cvxopt=use_cvxopt) # heuristics need to be adapted for prefix alignments # here we make the heuristics way less powerful h = h / (max_trace_length + 1.0) # 11/10/19: shall not a state for which we compute the exact heuristics be # by nature a trusted solution? tp = SearchTuple(curr.g + h, curr.g, h, curr.m, curr.p, curr.t, x, True) # 11/10/2019 (optimization ZA) heappushpop is slightly more efficient than pushing # and popping separately curr = heapq.heappushpop(open_set, tp) current_marking = curr.m # max allowed heuristics value (27/10/2019, due to the numerical instability of some of our solvers) if curr.h > lp_solver_factory.MAX_ALLOWED_HEURISTICS: continue # 12/10/2019: do it again, since the marking could be changed already_closed = current_marking in closed if already_closed: continue # 12/10/2019: the current marking can be equal to the final marking only if the heuristics # (underestimation of the remaining cost) is 0. Low-hanging fruits enab_trans = [ x for x in petri.semantics.enabled_transitions( sync_net, current_marking) ] if stop <= current_marking: #print(utils.__reconstruct_alignment(curr, visited, queued, traversed)) enab_trans = [ x for x in sync_net.transitions if x.sub_marking <= current_marking ] return current_marking closed.add(current_marking) visited += 1 possible_enabling_transitions = set() for p in current_marking: for t in p.ass_trans: possible_enabling_transitions.add(t) enabled_trans = [ t for t in possible_enabling_transitions if t.sub_marking <= current_marking ] trans_to_visit_with_cost = [ (t, cost_function[t]) for t in enabled_trans if not (curr.t is not None and utils.__is_log_move(curr.t, skip) and utils.__is_model_move(t, skip)) ] for t, cost in trans_to_visit_with_cost: traversed += 1 new_marking = utils.add_markings(current_marking, t.add_marking) if new_marking in closed: continue g = curr.g + cost queued += 1 h, x = utils.__derive_heuristic(incidence_matrix, cost_vec, curr.x, t, curr.h) # heuristics need to be adapted for prefix alignments # here we make the heuristics way less powerful h = h / (max_trace_length + 1.0) trustable = utils.__trust_solution(x) new_f = g + h tp = SearchTuple(new_f, g, h, new_marking, curr, t, x, trustable) heapq.heappush(open_set, tp)
def __search(sync_net, ini, fin, cost_function, skip, ret_tuple_as_trans_desc=False, max_align_time_trace=sys.maxsize): start_time = time.time() decorate_transitions_prepostset(sync_net) decorate_places_preset_trans(sync_net) incidence_matrix = inc_mat_construct(sync_net) ini_vec, fin_vec, cost_vec = utils.__vectorize_initial_final_cost( incidence_matrix, ini, fin, cost_function) closed = set() a_matrix = np.asmatrix(incidence_matrix.a_matrix).astype(np.float64) g_matrix = -np.eye(len(sync_net.transitions)) h_cvx = np.matrix(np.zeros(len(sync_net.transitions))).transpose() cost_vec = [x * 1.0 for x in cost_vec] use_cvxopt = False if lp_solver.DEFAULT_LP_SOLVER_VARIANT == lp_solver.CVXOPT_SOLVER_CUSTOM_ALIGN or lp_solver.DEFAULT_LP_SOLVER_VARIANT == lp_solver.CVXOPT_SOLVER_CUSTOM_ALIGN_ILP: use_cvxopt = True if use_cvxopt: # not available in the latest version of PM4Py from cvxopt import matrix a_matrix = matrix(a_matrix) g_matrix = matrix(g_matrix) h_cvx = matrix(h_cvx) cost_vec = matrix(cost_vec) h, x = utils.__compute_exact_heuristic_new_version( sync_net, a_matrix, h_cvx, g_matrix, cost_vec, incidence_matrix, ini, fin_vec, lp_solver.DEFAULT_LP_SOLVER_VARIANT, use_cvxopt=use_cvxopt) ini_state = utils.SearchTuple(0 + h, 0, h, ini, None, None, x, True) open_set = [ini_state] heapq.heapify(open_set) visited = 0 queued = 0 traversed = 0 lp_solved = 1 trans_empty_preset = set(t for t in sync_net.transitions if len(t.in_arcs) == 0) while not len(open_set) == 0: if (time.time() - start_time) > max_align_time_trace: return None curr = heapq.heappop(open_set) current_marking = curr.m while not curr.trust: if (time.time() - start_time) > max_align_time_trace: return None already_closed = current_marking in closed if already_closed: curr = heapq.heappop(open_set) current_marking = curr.m continue h, x = utils.__compute_exact_heuristic_new_version( sync_net, a_matrix, h_cvx, g_matrix, cost_vec, incidence_matrix, curr.m, fin_vec, lp_solver.DEFAULT_LP_SOLVER_VARIANT, use_cvxopt=use_cvxopt) lp_solved += 1 # 11/10/19: shall not a state for which we compute the exact heuristics be # by nature a trusted solution? tp = utils.SearchTuple(curr.g + h, curr.g, h, curr.m, curr.p, curr.t, x, True) # 11/10/2019 (optimization ZA) heappushpop is slightly more efficient than pushing # and popping separately curr = heapq.heappushpop(open_set, tp) current_marking = curr.m # max allowed heuristics value (27/10/2019, due to the numerical instability of some of our solvers) if curr.h > lp_solver.MAX_ALLOWED_HEURISTICS: continue # 12/10/2019: do it again, since the marking could be changed already_closed = current_marking in closed if already_closed: continue # 12/10/2019: the current marking can be equal to the final marking only if the heuristics # (underestimation of the remaining cost) is 0. Low-hanging fruits if curr.h < 0.01: if current_marking == fin: return utils.__reconstruct_alignment( curr, visited, queued, traversed, ret_tuple_as_trans_desc=ret_tuple_as_trans_desc, lp_solved=lp_solved) closed.add(current_marking) visited += 1 possible_enabling_transitions = copy(trans_empty_preset) for p in current_marking: for t in p.ass_trans: possible_enabling_transitions.add(t) enabled_trans = [ t for t in possible_enabling_transitions if t.sub_marking <= current_marking ] trans_to_visit_with_cost = [ (t, cost_function[t]) for t in enabled_trans if not (t is not None and utils.__is_log_move(t, skip) and utils.__is_model_move(t, skip)) ] for t, cost in trans_to_visit_with_cost: traversed += 1 new_marking = utils.add_markings(current_marking, t.add_marking) if new_marking in closed: continue g = curr.g + cost queued += 1 h, x = utils.__derive_heuristic(incidence_matrix, cost_vec, curr.x, t, curr.h) trustable = utils.__trust_solution(x) new_f = g + h tp = utils.SearchTuple(new_f, g, h, new_marking, curr, t, x, trustable) heapq.heappush(open_set, tp)
def __search(sync_net, ini, fin, cost_function, skip, ret_tuple_as_trans_desc=False, max_align_time_trace=sys.maxsize): start_time = time.time() decorate_transitions_prepostset(sync_net) decorate_places_preset_trans(sync_net) incidence_matrix = inc_mat_construct(sync_net) ini_vec, fin_vec, cost_vec = utils.__vectorize_initial_final_cost( incidence_matrix, ini, fin, cost_function) closed = set() heu_dict = {} heu_max_ind_dict = {} mtcgt_dict = {} parameters = {} parameters[marking_equation.Parameters.FULL_BOOTSTRAP_REQUIRED] = False parameters[marking_equation.Parameters.INCIDENCE_MATRIX] = incidence_matrix parameters[marking_equation.Parameters.COSTS] = cost_function visited = 0 queued = 0 traversed = 0 me = marking_equation.build(sync_net, ini, fin, parameters=parameters) h, x = me.solve() lp_solved = 1 # try to see if the firing sequence is already fine firing_sequence, reach_fm, explained_events = me.get_firing_sequence(x) if reach_fm: return __reconstruct_alignment( firing_sequence, h, visited, queued, traversed, ret_tuple_as_trans_desc=ret_tuple_as_trans_desc, lp_solved=lp_solved) mm, index = __get_model_marking_and_index(ini) __update_heu_dict(heu_dict, heu_max_ind_dict, mm, index, h, x, firing_sequence, incidence_matrix, cost_vec) ini_state = utils.TweakedSearchTuple(0 + h, 0, h, ini, None, None, x, True, False) open_set = [ini_state] heapq.heapify(open_set) trans_empty_preset = set(t for t in sync_net.transitions if len(t.in_arcs) == 0) while not len(open_set) == 0: if (time.time() - start_time) > max_align_time_trace: return None curr = heapq.heappop(open_set) current_marking = curr.m while not curr.trust: if (time.time() - start_time) > max_align_time_trace: return None already_closed = current_marking in closed if already_closed: curr = heapq.heappop(open_set) current_marking = curr.m continue if curr.t not in mtcgt_dict: lp_solved += 1 mtcgt = __min_total_cost_given_trans(me, ini, incidence_matrix, curr.t) mtcgt_dict[curr.t] = mtcgt else: mtcgt = mtcgt_dict[curr.t] h1 = max(mtcgt - curr.g, 0) if h1 > curr.h: tp = utils.TweakedSearchTuple(curr.g + h1, curr.g, h1, curr.m, curr.p, curr.t, curr.x, False, False) curr = heapq.heappushpop(open_set, tp) current_marking = curr.m continue mm, index = __get_model_marking_and_index(curr.m) h2, x2, trust2 = __get_heu_from_dict(heu_dict, heu_max_ind_dict, mm, index) if h2 is not None and h2 > curr.h: tp = utils.TweakedSearchTuple(curr.g + h2, curr.g, h2, curr.m, curr.p, curr.t, x2, trust2, False) curr = heapq.heappushpop(open_set, tp) current_marking = curr.m continue me.change_ini_vec(curr.m) h, x = me.solve() __update_heu_dict_specific_point(heu_dict, heu_max_ind_dict, mm, index, h, x) lp_solved += 1 tp = utils.TweakedSearchTuple(curr.g + h, curr.g, h, curr.m, curr.p, curr.t, x, True, True) curr = heapq.heappushpop(open_set, tp) current_marking = curr.m already_closed = current_marking in closed if already_closed: continue if curr.h < 0.01: if current_marking == fin: trans_list = __transitions_list_from_state(curr) return __reconstruct_alignment( trans_list, curr.f, visited, queued, traversed, ret_tuple_as_trans_desc=ret_tuple_as_trans_desc, lp_solved=lp_solved) if curr.virgin: # try to see if the firing sequence is already fine firing_sequence, reach_fm, explained_events = me.get_firing_sequence( curr.x) if reach_fm: trans_list = __transitions_list_from_state(curr) + list( firing_sequence) return __reconstruct_alignment( trans_list, curr.f, visited, queued, traversed, ret_tuple_as_trans_desc=ret_tuple_as_trans_desc, lp_solved=lp_solved) mm, index = __get_model_marking_and_index(curr.m) __update_heu_dict(heu_dict, heu_max_ind_dict, mm, index, h, x, firing_sequence, incidence_matrix, cost_vec) closed.add(current_marking) visited += 1 possible_enabling_transitions = copy(trans_empty_preset) for p in current_marking: for t in p.ass_trans: possible_enabling_transitions.add(t) enabled_trans = [ t for t in possible_enabling_transitions if t.sub_marking <= current_marking ] trans_to_visit_with_cost = [ (t, cost_function[t]) for t in enabled_trans if not (t is not None and utils.__is_log_move(t, skip) and utils.__is_model_move(t, skip)) ] for t, cost in trans_to_visit_with_cost: traversed += 1 new_marking = utils.add_markings(current_marking, t.add_marking) if new_marking in closed: continue g = curr.g + cost queued += 1 h, x = utils.__derive_heuristic(incidence_matrix, cost_vec, curr.x, t, curr.h) trust = utils.__trust_solution(x) mm, index = __get_model_marking_and_index(new_marking) if not trust: h2, x2, trust2 = __get_heu_from_dict(heu_dict, heu_max_ind_dict, mm, index) if h2 is not None and (h2 > h or trust2): h = h2 x = x2 trust = trust2 else: __update_heu_dict_specific_point(heu_dict, heu_max_ind_dict, mm, index, h, x) new_f = g + h tp = utils.TweakedSearchTuple(new_f, g, h, new_marking, curr, t, x, trust, False) heapq.heappush(open_set, tp)