def solve(Ws, Cs, Qs, ws, b, transitive_reduction=True,
svg_file="", lp_file="", mps_file="",
script=None, script_options=None, stats=None, verbose=None):
"""
Solve multiple-choice vector bin packing instances
using the method proposed in:
Brandao, F. and Pedroso, J. P. (2013). Multiple-choice Vector
Bin Packing: Arc-flow Formulation with Graph Compression. Technical Report
DCC-2013-13, Faculdade de Ciencias da Universidade do Porto, Universidade
do Porto, Portugal.
"""
assert script is not None
assert svg_file == "" or svg_file.endswith(".svg")
if stats is None and verbose is not None:
stats = verbose
nbtypes = len(Ws)
ndims = len(Ws[0])
ww = []
bb = []
itlabel = []
for i in range(len(ws)):
for j in range(len(ws[i])):
itlabel.append((i, j))
ww.append(ws[i][j])
bb.append(b[i])
LOSS = "L"
instances = [None] * nbtypes
graphs = [None] * nbtypes
Ss, Ts = [None] * nbtypes, [None] * nbtypes
for i in range(nbtypes):
bbi = bb[:]
for j in range(len(ww)):
if any(a > b for a, b in zip(ww[j], Ws[i])):
bbi[j] = 0
continue
symb = "G{0}:".format(i)
instances[i] = VBP(Ws[i], ww, bbi, verbose=False)
graphs[i] = AFG(instances[i], verbose=verbose).graph()
loss = graphs[i].LOSS
graphs[i].relabel(
lambda u: "{0}{1}".format(symb, u),
lambda lbl: lbl if lbl != loss else LOSS
)
Ss[i] = graphs[i].S
Ts[i] = graphs[i].Ts[0]
graphs[i].A.remove((Ts[i], Ss[i], LOSS))
if svg_file.endswith(".svg"):
try:
graphs[i].draw(
svg_file.replace(".svg", "{0}.svg".format(i+1)),
verbose=verbose
)
except Exception as e:
VPSolver.log(e, verbose)
S, T = "S", "T"
V = sum([g.V for g in graphs], [])
A = sum([g.A for g in graphs], [])
V += [S, T]
A += [(S, s, LOSS) for s in Ss]
A += [(t, T, LOSS) for t in Ts]
A += [(T, S, LOSS)]
graph = AFGraph(V, A, S, [T], LOSS)
if svg_file.endswith(".svg"):
try:
graph.draw(svg_file, verbose=verbose)
except Exception as e:
VPSolver.log(e, verbose)
adj = {u: [] for u in V}
for (u, v, i) in A:
adj[v].append((u, i))
VPSolver.log("Final compression steps:", verbose)
nv1, na1 = len(V), len(A)
VPSolver.log(" #V1: {0} #A1: {1}".format(nv1, na1), verbose=stats)
zero = tuple([0]*ndims)
def compress(u):
if u == S:
return zero
if u in newlbl:
return newlbl[u]
lbl = zero
for v, i in adj[u]:
wi = ww[i] if i != LOSS else zero
vlbl = compress(v)
lbl = tuple(max(lbl[d], vlbl[d]+wi[d]) for d in range(ndims))
newlbl[u] = lbl
return lbl
# Relabel the graph using the longest path from the source:
newlbl = {}
compress(T)
newlbl[S] = S
newlbl[T] = T
for t in Ts:
newlbl[t] = t
for u in newlbl:
if newlbl[u] == zero:
newlbl[u] = S
graph.relabel(lambda u: newlbl.get(u, u))
if transitive_reduction:
# Reduce graph size by connecting nodes to non-dominated targets only:
tadj = {}
for (u, v, i) in graph.A:
if isinstance(u, tuple):
if v in Ts:
if u not in tadj:
tadj[u] = []
tadj[u].append(Ts.index(v))
graph.A = [
(u, v, i) for (u, v, i) in graph.A if v not in Ts
]
V, A = set(graph.V), set(graph.A)
def fits(w1, w2):
return all(x <= y for x, y in zip(w1, w2))
dominatedby = {i: [] for i in range(nbtypes)}
for i in range(nbtypes):
for j in range(nbtypes):
if i < j and Ws[i] == Ws[j]:
dominatedby[i].append(j)
elif i != j and fits(Ws[i], Ws[j]):
dominatedby[i].append(j)
for v in tadj:
tgts = set(tadj[v])
for i in range(nbtypes):
if i not in tgts:
continue
for j in dominatedby[i]:
try:
tgts.remove(j)
except:
pass
for t in tgts:
A.add((v, Ts[t], LOSS))
for i in range(nbtypes):
tgts = set(dominatedby[i])
for j in dominatedby[i]:
for t in dominatedby[j]:
try:
tgts.remove(t)
except:
pass
for t in tgts:
A.add((Ts[i], Ts[t], LOSS))
graph.V, graph.A = V, A
# Relabel the graph with indices:
newlbl = {}
for u in graph.get_vertices_sorted():
if isinstance(u, tuple):
newlbl[u] = len(newlbl)+1
graph.relabel(lambda u: newlbl.get(u, u))
V, A = graph.V, graph.A
nv2, na2 = len(V), len(A)
VPSolver.log(" #V2: {0} #A2: {1}".format(nv2, na2), verbose=stats)
VPSolver.log(" #V2/#V1 = {0:.2f}".format(nv2/nv1), verbose=stats)
VPSolver.log(" #A2/#A1 = {0:.2f}".format(na2/na1), verbose=stats)
if svg_file.endswith(".svg"):
try:
graph.draw(svg_file.replace(".svg", ".final.svg"), verbose=verbose)
except Exception as e:
VPSolver.log(e, verbose)
# Remove redudant parallel arcs:
At = []
used = set()
for (u, v, i) in A:
k = itlabel[i][0] if i != LOSS else LOSS
if (u, v, k) not in used:
At.append((u, v, i))
used.add((u, v, k))
A = At
graph = AFGraph(V, A, S, [T], LOSS)
nv3, na3 = len(V), len(A)
VPSolver.log(" #V3: {0} #A3: {1}".format(nv3, na3), verbose=stats)
VPSolver.log(" #V3/#V1 = {0:.2f}".format(nv3/nv1), verbose=stats)
VPSolver.log(" #A3/#A1 = {0:.2f}".format(na3/na1), verbose=stats)
# Generate the model:
varl, cons = graph.get_flow_cons()
assocs = graph.get_assocs()
for i in range(len(b)):
lincomb = [
(var, 1)
for it, (j, t) in enumerate(itlabel) if j == i
for var in assocs[it]
]
if b[i] > 1:
cons.append((lincomb, ">=", b[i]))
else:
cons.append((lincomb, "=", b[i]))
model = Model()
ub = {}
for i in range(len(b)):
for var in assocs[i]:
ub[var] = b[i]
n = sum(b)
for i in range(nbtypes):
var = graph.vname(Ts[i], T, LOSS)
if Qs[i] == -1:
ub[var] = n
else:
ub[var] = min(Qs[i], n)
for var in varl:
# model.add_var(name=var, lb=0, vtype="I")
model.add_var(name=var, lb=0, ub=ub.get(var, None), vtype="I")
for lincomb, sign, rhs in cons:
model.add_con(lincomb, sign, rhs)
lincomb = [(graph.vname(Ts[i], T, LOSS), Cs[i]) for i in range(nbtypes)]
model.set_obj("min", lincomb)
model_file = VPSolver.new_tmp_file(".lp")
model.write(model_file)
if lp_file.endswith(".lp"):
model.write(lp_file)
VPSolver.log(".LP model successfully generated!", verbose)
if mps_file.endswith(".mps"):
model.write(mps_file)
VPSolver.log(".MPS model successfully generated!", verbose)
out, varvalues = VPSolver.script_wsol(script, model_file, verbose=verbose)
os.remove(model_file)
VPSolver.log("#V1: {0} #A1: {1}".format(nv1, na1), verbose)
VPSolver.log("#V2: {0} #A2: {1}".format(nv2, na2), verbose)
VPSolver.log("#V3: {0} #A3: {1}".format(nv3, na3), verbose)
VPSolver.log(
"#V3/#V1: {0:.2f} #A3/#A1: {1:.2f}".format(
nv3/nv1, na3/na1
),
verbose
)
labels = {}
for (u, v, i) in A:
if i != LOSS:
labels[u, v, i] = [itlabel[i]]
lst_sol = []
graph.set_flow(varvalues)
graph.set_labels(labels)
for i in range(nbtypes):
lst_sol.append(graph.extract_solution(S, "<-", Ts[i]))
validate_solution(lst_sol, nbtypes, ndims, Ws, ws, b)
c1 = sum(sum(r for r, patt in lst_sol[i])*Cs[i] for i in range(nbtypes))
c2 = sum(
varvalues.get(graph.vname(Ts[i], T, LOSS), 0) * Cs[i]
for i in range(nbtypes)
)
assert c1 == c2
return c1, lst_sol
def solve(
Ws, Cs, Qs, ws, b, svg_file="", lp_file="", mps_file="",
script="vpsolver_glpk.sh", verbose=False):
"""
Solves multiple-choice vector bin packing instances
using the method proposed in:
Brandao, F. and Pedroso, J. P. (2013). Multiple-choice Vector
Bin Packing: Arc-flow Formulation with Graph Compression. Technical Report
DCC-2013-13, Faculdade de Ciencias da Universidade do Porto, Universidade
do Porto, Portugal.
"""
assert svg_file == "" or svg_file.endswith(".svg")
nbtypes = len(Ws)
ndims = len(Ws[0])
ww = []
bb = []
itlabel = []
for i in range(len(ws)):
for j in range(len(ws[i])):
itlabel.append((i, j))
ww.append(ws[i][j])
bb.append(b[i])
instances = [None] * nbtypes
graphs = [None] * nbtypes
Ss, Ts = [None] * nbtypes, [None] * nbtypes
for i in range(nbtypes):
bbi = bb[:]
for j in range(len(ww)):
if any(a > b for a, b in zip(ww[j], Ws[i])):
bbi[j] = 0
continue
symb = "G{0}".format(i)
instances[i] = VBP(Ws[i], ww, bbi, verbose=False)
graphs[i] = AFG(instances[i], verbose=verbose).graph()
graphs[i].relabel(lambda u: "{0}{1}".format(symb, u))
Ss[i] = symb+"S"
Ts[i] = symb+"T"
if svg_file.endswith(".svg"):
try:
graphs[i].draw(svg_file.replace(".svg", "{0}.svg".format(i)))
except Exception as e:
VPSolver.log(e, verbose)
V = sum([g.V for g in graphs], [])
A = sum([g.A for g in graphs], [])
V += ["S", "T"]
A += [("S", s, "L") for s in Ss]
A += [(t, "T", "L") for t in Ts]
A += [("T", "S", "L")]
graph = AFGraph(V, A, "S", "T")
if svg_file.endswith(".svg"):
try:
graph.draw(svg_file, ignore=[("T", "S")])
except Exception as e:
VPSolver.log(e, verbose)
adj = {u: [] for u in V}
for (u, v, i) in A:
adj[v].append((u, i))
S, T = "S", "T"
newlbl = {}
def compress(u):
if u == S:
return [0]*ndims
if u in newlbl:
return newlbl[u]
def itemw(lbl):
if isinstance(lbl, int) and lbl < len(ww):
return ww[lbl]
else:
return [0]*ndims
lbl = [0]*ndims
for v, i in adj[u]:
wi = itemw(i)
vlbl = compress(v)
for d in range(ndims):
lbl[d] = max(lbl[d], vlbl[d]+wi[d])
newlbl[u] = lbl
return lbl
compress(T)
revlbl = {tuple(x): [] for x in newlbl.values()}
for u in newlbl:
if u not in Ts+[T]:
revlbl[tuple(newlbl[u])].append(u)
revlbl2 = {}
for i, x in enumerate(sorted(revlbl)):
if sum(x) != 0:
revlbl2["V{0}".format(i)] = revlbl[x]
else:
revlbl2[S] = revlbl[x]
vlbl = {}
for lbl in revlbl2:
for v in revlbl2[lbl]:
vlbl[v] = lbl
assert set([v for v in V if v not in vlbl]) == set([S, T]+Ts)
VPSolver.log("Final compression steps:", verbose)
nv1, na1 = len(V), len(A)
VPSolver.log(" #V1: {0} #A1: {1}".format(nv1, na1), verbose)
graph.relabel(lambda u: vlbl.get(u, u))
V, A = graph.V, graph.A
nv2, na2 = len(V), len(A)
VPSolver.log(" #V2: {0} #A2: {1}".format(nv2, na2), verbose)
VPSolver.log(" #V2/#V1 = {0:.2f}".format(nv2/nv1), verbose)
VPSolver.log(" #A2/#A1 = {0:.2f}".format(na2/na1), verbose)
if svg_file.endswith(".svg"):
try:
graph.draw(
svg_file.replace(".svg", ".final.svg"), ignore=[("T", "S")]
)
except Exception as e:
VPSolver.log(e, verbose)
# remove redudant parallel arcs
At = []
used = set()
for (u, v, i) in A:
k = itlabel[i][0] if (isinstance(i, int) and i < len(itlabel)) else "L"
if (u, v, k) not in used:
At.append((u, v, i))
used.add((u, v, k))
A = At
graph = AFGraph(V, A, "S", "T")
nv3, na3 = len(V), len(A)
VPSolver.log(" #V3: {0} #A3: {1}".format(nv3, na3), verbose)
VPSolver.log(" #V3/#V1 = {0:.2f}".format(nv3/nv1), verbose)
VPSolver.log(" #A3/#A1 = {0:.2f}".format(na3/na1), verbose)
varl, cons = graph.get_flow_cons()
assocs = graph.get_assocs()
for i in range(len(b)):
lincomb = [
(var, 1)
for it, (j, t) in enumerate(itlabel) if j == i
for var in assocs[it]
]
# cons.append((lincomb, ">=", b[i]))
if b[i] > 1:
cons.append((lincomb, ">=", b[i]))
else:
cons.append((lincomb, "=", b[i]))
model = Model()
ub = {}
for i in range(len(b)):
for var in assocs[i]:
ub[var] = b[i]
n = sum(b)
for i in range(nbtypes):
var = graph.vname(Ts[i], "T", "L")
ub[var] = min(Qs[i], n)
for var in varl:
# model.add_var(name=var, lb=0, vtype="I")
model.add_var(name=var, lb=0, ub=ub.get(var, None), vtype="I")
for lincomb, sign, rhs in cons:
model.add_con(lincomb, sign, rhs)
lincomb = [(graph.vname(Ts[i], "T", "L"), Cs[i]) for i in range(nbtypes)]
model.set_obj("min", lincomb)
model_file = VPSolver.new_tmp_file(".lp")
model.write(model_file)
if lp_file.endswith(".lp"):
model.write(lp_file)
VPSolver.log(".LP model successfully generated!", verbose)
if mps_file.endswith(".mps"):
model.write(mps_file)
VPSolver.log(".MPS model successfully generated!", verbose)
out, varvalues = VPSolver.script_wsol(script, model_file, verbose=verbose)
os.remove(model_file)
VPSolver.log("#V1: {0} #A1: {1}".format(nv1, na1), verbose)
VPSolver.log("#V2: {0} #A2: {1}".format(nv2, na2), verbose)
VPSolver.log("#V3: {0} #A3: {1}".format(nv3, na3), verbose)
VPSolver.log(
"#V3/#V1: {0:.2f} #A3/#A1: {1:.2f}".format(
nv3/nv1, na3/na1
),
verbose
)
labels = {}
for (u, v, i) in A:
if isinstance(i, int) and i < len(itlabel):
labels[u, v, i] = [itlabel[i]]
lst_sol = []
graph.set_flow(varvalues)
graph.set_labels(labels)
for i in range(nbtypes):
lst_sol.append(graph.extract_solution("S", "<-", Ts[i]))
assert graph.validate_solution(lst_sol, nbtypes, ndims, Ws, ws, b)
c1 = sum(sum(r for r, patt in lst_sol[i])*Cs[i] for i in range(nbtypes))
c2 = sum(
varvalues.get(graph.vname(Ts[i], "T", "L"), 0) * Cs[i]
for i in range(nbtypes)
)
assert c1 == c2
return c1, lst_sol
def solve(Ws,
Cs,
Qs,
ws,
b,
transitive_reduction=True,
svg_file="",
lp_file="",
mps_file="",
script=None,
script_options=None,
stats=None,
verbose=None):
"""
Solve multiple-choice vector bin packing instances
using the method proposed in:
Brandao, F. and Pedroso, J. P. (2013). Multiple-choice Vector
Bin Packing: Arc-flow Formulation with Graph Compression. Technical Report
DCC-2013-13, Faculdade de Ciencias da Universidade do Porto, Universidade
do Porto, Portugal.
"""
assert script is not None
assert svg_file == "" or svg_file.endswith(".svg")
if stats is None and verbose is not None:
stats = verbose
nbtypes = len(Ws)
ndims = len(Ws[0])
ww = []
bb = []
itlabel = []
for i in range(len(ws)):
for j in range(len(ws[i])):
itlabel.append((i, j))
ww.append(ws[i][j])
bb.append(b[i])
LOSS = "L"
instances = [None] * nbtypes
graphs = [None] * nbtypes
Ss, Ts = [None] * nbtypes, [None] * nbtypes
for i in range(nbtypes):
bbi = bb[:]
for j in range(len(ww)):
if any(a > b for a, b in zip(ww[j], Ws[i])):
bbi[j] = 0
continue
symb = "G{0}:".format(i)
instances[i] = VBP(Ws[i], ww, bbi, verbose=False)
graphs[i] = AFG(instances[i], verbose=verbose).graph()
loss = graphs[i].LOSS
graphs[i].relabel(lambda u: "{0}{1}".format(symb, u), lambda lbl: lbl
if lbl != loss else LOSS)
Ss[i] = graphs[i].S
Ts[i] = graphs[i].Ts[0]
graphs[i].A.remove((Ts[i], Ss[i], LOSS))
if svg_file.endswith(".svg"):
try:
graphs[i].draw(svg_file.replace(".svg",
"{0}.svg".format(i + 1)),
verbose=verbose)
except Exception as e:
VPSolver.log(e, verbose)
S, T = "S", "T"
V = sum([g.V for g in graphs], [])
A = sum([g.A for g in graphs], [])
V += [S, T]
A += [(S, s, LOSS) for s in Ss]
A += [(t, T, LOSS) for t in Ts]
A += [(T, S, LOSS)]
graph = AFGraph(V, A, S, [T], LOSS)
if svg_file.endswith(".svg"):
try:
graph.draw(svg_file, verbose=verbose)
except Exception as e:
VPSolver.log(e, verbose)
adj = {u: [] for u in V}
for (u, v, i) in A:
adj[v].append((u, i))
VPSolver.log("Final compression steps:", verbose)
nv1, na1 = len(V), len(A)
VPSolver.log(" #V1: {0} #A1: {1}".format(nv1, na1), verbose=stats)
zero = tuple([0] * ndims)
def compress(u):
if u == S:
return zero
if u in newlbl:
return newlbl[u]
lbl = zero
for v, i in adj[u]:
wi = ww[i] if i != LOSS else zero
vlbl = compress(v)
lbl = tuple(max(lbl[d], vlbl[d] + wi[d]) for d in range(ndims))
newlbl[u] = lbl
return lbl
# Relabel the graph using the longest path from the source:
newlbl = {}
compress(T)
newlbl[S] = S
newlbl[T] = T
for t in Ts:
newlbl[t] = t
for u in newlbl:
if newlbl[u] == zero:
newlbl[u] = S
graph.relabel(lambda u: newlbl.get(u, u))
if transitive_reduction:
# Reduce graph size by connecting nodes to non-dominated targets only:
tadj = {}
for (u, v, i) in graph.A:
if isinstance(u, tuple):
if v in Ts:
if u not in tadj:
tadj[u] = []
tadj[u].append(Ts.index(v))
graph.A = [(u, v, i) for (u, v, i) in graph.A if v not in Ts]
V, A = set(graph.V), set(graph.A)
def fits(w1, w2):
return all(x <= y for x, y in zip(w1, w2))
dominatedby = {i: [] for i in range(nbtypes)}
for i in range(nbtypes):
for j in range(nbtypes):
if i < j and Ws[i] == Ws[j]:
dominatedby[i].append(j)
elif i != j and fits(Ws[i], Ws[j]):
dominatedby[i].append(j)
for v in tadj:
tgts = set(tadj[v])
for i in range(nbtypes):
if i not in tgts:
continue
for j in dominatedby[i]:
try:
tgts.remove(j)
except:
pass
for t in tgts:
A.add((v, Ts[t], LOSS))
for i in range(nbtypes):
tgts = set(dominatedby[i])
for j in dominatedby[i]:
for t in dominatedby[j]:
try:
tgts.remove(t)
except:
pass
for t in tgts:
A.add((Ts[i], Ts[t], LOSS))
graph.V, graph.A = V, A
# Relabel the graph with indices:
newlbl = {}
for u in graph.get_vertices_sorted():
if isinstance(u, tuple):
newlbl[u] = len(newlbl) + 1
graph.relabel(lambda u: newlbl.get(u, u))
V, A = graph.V, graph.A
nv2, na2 = len(V), len(A)
VPSolver.log(" #V2: {0} #A2: {1}".format(nv2, na2), verbose=stats)
VPSolver.log(" #V2/#V1 = {0:.2f}".format(nv2 / nv1), verbose=stats)
VPSolver.log(" #A2/#A1 = {0:.2f}".format(na2 / na1), verbose=stats)
if svg_file.endswith(".svg"):
try:
graph.draw(svg_file.replace(".svg", ".final.svg"), verbose=verbose)
except Exception as e:
VPSolver.log(e, verbose)
# Remove redudant parallel arcs:
At = []
used = set()
for (u, v, i) in A:
k = itlabel[i][0] if i != LOSS else LOSS
if (u, v, k) not in used:
At.append((u, v, i))
used.add((u, v, k))
A = At
graph = AFGraph(V, A, S, [T], LOSS)
nv3, na3 = len(V), len(A)
VPSolver.log(" #V3: {0} #A3: {1}".format(nv3, na3), verbose=stats)
VPSolver.log(" #V3/#V1 = {0:.2f}".format(nv3 / nv1), verbose=stats)
VPSolver.log(" #A3/#A1 = {0:.2f}".format(na3 / na1), verbose=stats)
# Generate the model:
varl, cons = graph.get_flow_cons()
assocs = graph.get_assocs()
for i in range(len(b)):
lincomb = [(var, 1) for it, (j, t) in enumerate(itlabel) if j == i
for var in assocs[it]]
if b[i] > 1:
cons.append((lincomb, ">=", b[i]))
else:
cons.append((lincomb, "=", b[i]))
model = Model()
ub = {}
for i in range(len(b)):
for var in assocs[i]:
ub[var] = b[i]
n = sum(b)
for i in range(nbtypes):
var = graph.vname(Ts[i], T, LOSS)
if Qs[i] == -1:
ub[var] = n
else:
ub[var] = min(Qs[i], n)
for var in varl:
# model.add_var(name=var, lb=0, vtype="I")
model.add_var(name=var, lb=0, ub=ub.get(var, None), vtype="I")
for lincomb, sign, rhs in cons:
model.add_con(lincomb, sign, rhs)
lincomb = [(graph.vname(Ts[i], T, LOSS), Cs[i]) for i in range(nbtypes)]
model.set_obj("min", lincomb)
model_file = VPSolver.new_tmp_file(".lp")
model.write(model_file)
if lp_file.endswith(".lp"):
model.write(lp_file)
VPSolver.log(".LP model successfully generated!", verbose)
if mps_file.endswith(".mps"):
model.write(mps_file)
VPSolver.log(".MPS model successfully generated!", verbose)
out, varvalues = VPSolver.script_wsol(script, model_file, verbose=verbose)
os.remove(model_file)
VPSolver.log("#V1: {0} #A1: {1}".format(nv1, na1), verbose)
VPSolver.log("#V2: {0} #A2: {1}".format(nv2, na2), verbose)
VPSolver.log("#V3: {0} #A3: {1}".format(nv3, na3), verbose)
VPSolver.log(
"#V3/#V1: {0:.2f} #A3/#A1: {1:.2f}".format(nv3 / nv1, na3 / na1),
verbose)
labels = {}
for (u, v, i) in A:
if i != LOSS:
labels[u, v, i] = [itlabel[i]]
lst_sol = []
graph.set_flow(varvalues)
graph.set_labels(labels)
for i in range(nbtypes):
lst_sol.append(graph.extract_solution(S, "<-", Ts[i]))
validate_solution(lst_sol, nbtypes, ndims, Ws, ws, b)
c1 = sum(sum(r for r, patt in lst_sol[i]) * Cs[i] for i in range(nbtypes))
c2 = sum(
varvalues.get(graph.vname(Ts[i], T, LOSS), 0) * Cs[i]
for i in range(nbtypes))
assert c1 == c2
return c1, lst_sol
def test_model():
"""Test model."""
from pympl import Model, Tools, glpkutils
os.chdir(os.path.dirname(__file__) or os.curdir)
model = Model()
values = [15, 10, 9, 5]
weights = [1, 5, 3, 4]
xvars = []
for i in range(len(values)):
var = model.add_var(lb=0, ub=1, vtype="I")
xvars.append(var)
profit = model.add_var(lb=0, vtype="C")
model.add_con([(x, w) for x, w in zip(xvars, weights)], "<=", 8)
model.add_con(profit, "=", [(x, v) for x, v in zip(xvars, values)])
model.set_obj("max", profit)
lp_out = "tmp/model.lp"
mps_out = "tmp/model.mps"
mod_out = "tmp/model.mod"
model.write(lp_out)
model.write(mps_out)
model.write(mod_out)
out, varvalues = Tools.script("glpk_wrapper.sh", lp_out, verbose=True)
assert varvalues[profit] == 29
out, varvalues = Tools.script("glpk_wrapper.sh",
mps_out,
options="--max",
verbose=True)
assert varvalues[profit] == 29
glpkutils.mod2lp(mod_out, lp_out, verbose=True)
out, varvalues = Tools.script("glpk_wrapper.sh", lp_out, verbose=True)
assert varvalues[profit] == 29
def test_model():
"""Test model."""
from pympl import Model, Tools, glpkutils
os.chdir(os.path.dirname(__file__) or os.curdir)
model = Model()
values = [15, 10, 9, 5]
weights = [1, 5, 3, 4]
xvars = []
for i in range(len(values)):
var = model.add_var(lb=0, ub=1, vtype="I")
xvars.append(var)
profit = model.add_var(lb=0, vtype="C")
model.add_con([(x, w) for x, w in zip(xvars, weights)], "<=", 8)
model.add_con(profit, "=", [(x, v) for x, v in zip(xvars, values)])
model.set_obj("max", profit)
lp_out = "tmp/model.lp"
mps_out = "tmp/model.mps"
mod_out = "tmp/model.mod"
model.write(lp_out)
model.write(mps_out)
model.write(mod_out)
out, varvalues = Tools.script("glpk_wrapper.sh", lp_out, verbose=True)
assert varvalues[profit] == 29
out, varvalues = Tools.script(
"glpk_wrapper.sh", mps_out, options="--max", verbose=True
)
assert varvalues[profit] == 29
glpkutils.mod2lp(mod_out, lp_out, verbose=True)
out, varvalues = Tools.script("glpk_wrapper.sh", lp_out, verbose=True)
assert varvalues[profit] == 29