def main():
for g_input in gen_benchmark_digraphs():
# Giving up on Problem 10 after 2 hours:
# 0 0 11.00000 0 402 12.00000 11.00000 8.33% - 7981s
if g_input.graph['name'] == 'Problem 10 (opt=12)':
continue
solve_problem(g_input)
def main():
setParam("LogFile", "/tmp/gurobi.log")
for g in gen_benchmark_digraphs():
success, edgelist_per_cycle = get_all_cycles(g, cutoff=10000000)
if not success:
continue # too many simple cycles
solve_cm(g, edgelist_per_cycle)
def main():
#test_on_Jaconbsen_with_50_stages()
#return
start = time()
for g_input in gen_benchmark_digraphs():
solve_problem(g_input)
test_on_Jaconbsen_with_50_stages()
end = time()
print('Overall solution time: {0:0.1f} s'.format(end - start))
def main():
setParam("LogFile", "/tmp/gurobi.log")
for g in gen_benchmark_digraphs():
#if g.graph['name'] == 'Problem 10 (opt=12)':
#if g.graph['name'] != 'Problem 8 (opt=5)':
#if g.graph['name'] != 'Subproblem 8 (opt=3)':
# continue
STATISICS.clear()
# Problem 10
#-----------
# Sparse cut
#g.remove_edge(38, 36)
#g.remove_edge(39, 34)
#-----------
# Peeling off the DAG of the bypasses
#dag_connecting = [ (11,4), (17,16), (1,20), (26,25), (12,29), (35,34),
# (21,38), (44,43), (30,48), (54,53), (39,57) ]
#for e in dag_connecting:
# g.remove_edge(*e)
#-----------
# Breaking the small bypasses
#bypass_breakers = [ (17,11), (26,20), (35,29), (44,38), (54,48) ]
#for e in bypass_breakers:
# g.remove_edge(*e)
#-----------
# Problem 8
#g.remove_edge(4,6)
#g.remove_edge(24,16)
#g.remove_edge(17,22)
#g.remove_edge(21,20)
#g.remove_edge(2,6)
#g.remove_edge(24,26)
#iteratively_remove_runs_and_bypasses(g)
#g.remove_edge(2,17)
#g.remove_edge(8,27)
#g.remove_edge(27,7)
#g.remove_edge(8,14) # These latter two are insane but still OK
#g.remove_edge(18,21)
#plot(g, prog='sfdp')
#-----
#from utils import deserialize
#g = deserialize('data/JacobsenILOSimpBounds_as_DAG.pkl.gz')
#orig_input = deserialize('data/JacobsenILOSimpBounds_as_DAG.pkl.gz')
# And toggle comment on the # <- lines !!!
#-----
orig_input = deepcopy(g) # <-
simplify(g)
print('Info on the original input:')
print(orig_input.graph['name']) # <-
print('Nodes:', orig_input.number_of_nodes())
print('Edges:', orig_input.number_of_edges())
print('Loops:', len(list(nx.simple_cycles(orig_input)))) # <-
if 'max cycles' in STATISICS:
print('Loop budget:', STATISICS['max cycles'])
print('Cutoff:', STATISICS['max cutoff'])
plot(orig_input, prog='sfdp')
def run_tests():
for g, hints in gen_testgraphs():
plot(g, prog='sfdp')
run_elimination(g, hints)
# Comparing my block ordering to tearing (comparing apples to oranges).
for g in gen_benchmark_digraphs():
if g.number_of_selfloops():
print('Skipping it, as this digraph has self-loops\n')
continue
plot(g, prog='sfdp')
info(g)
run_elimination(g)
def real_main(use_min_degree):
#pname = 'JacobsenShortestSimpBounds'
pname = 'JacobsenILOSimpBounds'
g, eqs, forbidden = read_bipartite_graph(pname)
#feasible_solution = read_feasible_solution(pname, g, eqs, forbidden)
#
solve_problem(g, eqs, forbidden, use_min_degree)
#
for g, eqs, forbidden in gen_testproblems():
info_on_bipartite_graph(g, eqs, forbidden, log=log)
solve_problem(g, eqs, forbidden, use_min_degree)
#
# A bipartite graph, triggering poor performance
dig = next(gen_benchmark_digraphs())
for n in dig.nodes():
if n > 20:
dig.remove_node(n)
g, eqs, forbidden = digraph_to_undirected_bipartite(dig)
solve_problem(g, eqs, forbidden)
def main():
from benchmarks import gen_benchmark_digraphs
for g in gen_benchmark_digraphs():
info(g)
run_mfes_heuristic(g, try_one_cut=False, is_labeled=True)