Segmentation fault (core dumped)

Hi, I use Python to write a gurobi solver on a Linux system. When the problem size is small, it can be solved normally, but when the problem size increases, it reports an error: Segmentation fault. When I use the gdb tool to debug, the following message appears:

here is a part of my code

P_range = [i + 1 for i in range(len(self.LAYERS))]
S_range = [i for i in range(len(self.LAYERS))]
E_range = [i + 1 for i in range(len(self.LAYERS))]
W_range = [i for i in range(self.X+1)]
H_range = [i for i in range(self.Y+1)]
D_range = [0, 1]
self.Latency_Cache = np.load(os.path.join(self.result_path, self.name, 'Latency_Cache.npy'))
for S in S_range:
    for E in E_range:
        for W in W_range:
            for H in H_range:
                if not np.isnan(self.Latency_Cache[S][E][W][H]):
                    S_TEMP = S
                    E_TEMP = E
                    while (S_TEMP >= self.repeat_bottom and E_TEMP <= self.repeat_up):
                        self.Latency_Cache[S_TEMP + 1][E_TEMP + 1][W][H] = self.Latency_Cache[S_TEMP][E_TEMP][W][
                            H]
                        self.Latency_Cache[S_TEMP + 1][E_TEMP + 1][H][W] = self.Latency_Cache[S_TEMP][E_TEMP][W][
                            H]
                        S_TEMP = S_TEMP + 1
                        E_TEMP = E_TEMP + 1
                if W * H == 0:
                    self.Latency_Cache[S][E][W][H] == 1000
                if E-S == 0:
                    self.Latency_Cache[S][E][W][H] == 0
                if (self.Latency_Cache[S][E][W][H] == np.inf) or np.isnan(self.Latency_Cache[S][E][W][H]):
                    self.Latency_Cache[S][E][W][H] = 1000
print(self.Latency_Cache[0][24][16][16])


e = 0.001
M = 100 + e
model = grb.Model()
X = []
for S in S_range:
    X.append(model.addVars(P_range,  E_range[S:],vtype=grb.GRB.BINARY,name=f'X[{S}]'))

Y = model.addVars(P_range, D_range, W_range, H_range, vtype=grb.GRB.BINARY,name='Y')
Z = []#[S][E,W,H]
for S in S_range:
    Z.append(model.addVars(E_range[S:],W_range, H_range,vtype=grb.GRB.BINARY,name=f'Z[{S}]'))


SpVar = model.addVars(P_range, lb = 0, ub = S_range[-1], vtype=grb.GRB.INTEGER, name='SpVar')
EpVar = model.addVars(P_range, lb = 0, ub = E_range[-1], vtype=grb.GRB.INTEGER, name='EpVar')
maxEp = model.addVar( lb = 0, ub = E_range[-1], vtype=grb.GRB.INTEGER, name='maxEp')
minSp = model.addVar( lb = 0, ub = S_range[-1], vtype=grb.GRB.INTEGER, name='minSp')
CpVar = model.addVars(P_range, lb=W_range[0], ub=W_range[-1], vtype=grb.GRB.INTEGER, name='CpVar')
RpVar = model.addVars(P_range, lb=H_range[0], ub=H_range[-1], vtype=grb.GRB.INTEGER, name='RpVar')
maxCp = model.addVar(lb=0, ub=W_range[-1], vtype=grb.GRB.INTEGER, name='maxCp')
maxRp = model.addVar(lb=0, ub=H_range[-1], vtype=grb.GRB.INTEGER, name='maxRp')

PpXVar = model.addVars( P_range,vtype=grb.GRB.BINARY,name = 'PpXVar')
PpYVar = model.addVars(P_range, vtype=grb.GRB.BINARY, name='PpYVar')
TPpYVar = model.addVars(P_range, vtype=grb.GRB.BINARY, name='TPpYVar')
#constr_1_b = model.addVars(P_range[1:],vtype=grb.GRB.BINARY,name='constr_1_b')
AUX_OBJ = model.addVar(vtype=grb.GRB.CONTINUOUS,name='AUX_OBJ')
DpVar = model.addVars(P_range[1:],vtype=grb.GRB.BINARY,name= "DpVar")
Cp=[]
Rp=[]
Wp=[]
Hp=[]
Dp=[]
Sp=[]
Ep=[]
PpX=[]
PpY=[]
TPpY = []
expr_C = grb.LinExpr()
expr_R = grb.LinExpr()
expr_D = grb.LinExpr()
expr_W = grb.LinExpr()
expr_H = grb.LinExpr()
expr_TPY = grb.LinExpr()
expr_PY = grb.LinExpr()
expr_allP = grb.LinExpr()
for D in D_range:
    for W in W_range:
        for H in H_range:
            expr_C += Y[1, D, W, H] * W
            expr_R += Y[1, D, W, H] * H
            expr_D += Y[1, D, W, H] * D
            expr_W += Y[1, D, W, H] * W
            expr_H += Y[1, D, W, H] * H
            expr_PY += Y[1, D, W, H]
            if W * H > 0:
                expr_TPY += Y[1, D, W, H]
current_expr_C = expr_C.copy()
current_expr_R = expr_R.copy()
Dp.append(expr_D)
Cp.append(current_expr_C)
Rp.append(current_expr_R)
Wp.append(expr_W)
Hp.append(expr_H)
PpY.append(expr_PY)
TPpY.append(expr_TPY)
for P in P_range[1:]:
    expr_D = grb.LinExpr()
    expr_W = grb.LinExpr()
    expr_H = grb.LinExpr()
    expr_PY = grb.LinExpr()
    expr_TPY = grb.LinExpr()
    for D in D_range:
        for W in W_range:
            for H in H_range:
                if D == 0:
                    expr_C += Y[P,D,W,H]*W
                else:
                    expr_R += Y[P,D,W,H]*H
                current_expr_C = expr_C.copy()
                current_expr_R = expr_R.copy()#
                expr_D += Y[P, D, W, H] * D
                expr_W += Y[P, D, W, H] * W
                expr_H += Y[P, D, W, H] * H
                expr_PY += Y[P, D, W, H]
                if W*H > 0:
                    expr_TPY += Y[P, D, W, H]
    Dp.append(expr_D)
    Cp.append(current_expr_C)
    Rp.append(current_expr_R)
    Wp.append(expr_W)
    Hp.append(expr_H)
    PpY.append(expr_PY)
    TPpY.append(expr_TPY)
for P in P_range:
    expr_S = grb.LinExpr()
    expr_E = grb.LinExpr()
    expr_PX = grb.LinExpr()
    for S in S_range:
        for E in E_range[S:]:
            expr_S += X[S][P, E] * S
            expr_E += X[S][P, E] * E
            expr_PX += X[S][P, E]
            expr_allP += X[S][P, E]
    Sp.append(expr_S)
    Ep.append(expr_E)
    PpX.append(expr_PX)



model.addConstrs((DpVar[p] == Dp[p-1] for p in P_range[1:]),name="constr_1" )


model.addConstrs((Wp[p - 1] + M*(1 - DpVar[P]) >= Cp[p - 2] for p in P_range[1:]), name="constr_1_3_1")
model.addConstrs((Wp[p - 1] - M*(1 - DpVar[P]) <= Cp[p - 2] for p in P_range[1:]), name="constr_1_3_2")

model.addConstrs((Hp[p - 1] + M*DpVar[P] >= Rp[p - 2] for p in P_range[1:]), name="constr_1_3_3")
model.addConstrs((Hp[p - 1] - M*DpVar[P] <= Rp[p - 2] for p in P_range[1:]), name="constr_1_3_4")


model.addConstrs((PpXVar[p] == PpX[p - 1] for p in P_range),name = "constr_2_2")
model.addConstrs((EpVar[p] == Ep[p - 1] for p in P_range),name = "constr_2_3")
model.addConstrs((SpVar[p] == Sp[p - 1] for p in P_range),name = "constr_2_4")
model.addConstrs(((PpXVar[p] == 1) >> (SpVar[p] + e  <= EpVar[p])  for p in P_range), name = "constr_2_1")


model.addConstrs(((PpXVar[p] == 1) >> (EpVar[p - 1] == SpVar[p]) for p in P_range[1:]), name="constr_3_1")


model.addConstr(maxEp == max_([EpVar[p] for p in P_range]), name="constr_4_2")
#model.addConstrs((maxEp >= EpVar[p] for p in P_range), name="constr_4_2")

model.addConstr(len(self.LAYERS) == maxEp ,name = "const_4_3")


#model.addConstr(minSp == min_([SpVar[p] for p in P_range]), name="constr_5_2")
model.addConstr(0 == SpVar[P_range[0]], name="const_5_1")

model.addConstr((PpXVar[P_range[0]] == 1 ), name="constr_5_2")


model.addConstrs((PpYVar[p] == PpY[p - 1] for p in P_range), name="constr_6_0")
model.addConstrs((PpXVar[p - 1] >= PpXVar[p]  for p in P_range[1:]), name="constr_6_1")
model.addConstrs((PpXVar[p] <= 1 for p in P_range), name="constr_6_2")
model.addConstrs((PpYVar[p] == 1 for p in P_range), name="constr_6_3")


model.addConstr(expr_allP >= 1, name = "const_7")

model.addConstrs((TPpYVar[p] == TPpY[p - 1] for p in P_range), name="constr_8_0")
model.addConstrs((PpXVar[p] == TPpYVar[p] for p in P_range), name="constr_8_1")

model.addConstrs((CpVar[p] == Cp[p - 1] for p in P_range), name="const_9_1")
model.addConstr(maxCp == max_([CpVar[p] for p in P_range]), name="const_9_2")

model.addConstr(self.X == maxCp, name="const_9_3")

model.addConstrs((RpVar[p] == Rp[p - 1] for p in P_range), name="const_9_3")
model.addConstr(maxRp == max_([RpVar[p] for p in P_range]), name="const_9_4")
for S in S_range:
    for E in E_range[S:]:
        for W in W_range:
            for H in H_range:
                expr_temp = grb.QuadExpr()
                for P in P_range:
                    for D in D_range:
                        expr_temp += X[S][P,E] * Y[P, D, W, H]
                model.addConstr(Z[S][E, W, H] == expr_temp, name=f"constr_10_1_{S, E, W, H}")

model.addConstrs((expr_allP*Z[S][E, W, H]*self.Latency_Cache[S][E][W][H]<=self.max_latency for S in S_range
                  for E in E_range[S:] for W in W_range for H in H_range),name = "constr_11")
model.update()

model.addConstrs((AUX_OBJ>=Z[S][E, W, H]*self.Latency_Cache[S][E][W][H] for S in S_range
                  for E in E_range[S:] for W in W_range for H in H_range),name = "constr_obj")
model.setObjective(AUX_OBJ,sense=grb.GRB.MINIMIZE)
model.update()
model.optimize()

正式なコメント

Simranjit Kaur

Gurobi Staff

2026年11月13日 01:28

This post is more than two years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum, or try Gurobot, our chatbot interface offering instant, expert-level support.

YJ Z

2024年03月18日 08:26

When S_range =range(12), E_range=range(1,13),W_range = range(17),H_range = range(17), it can be executed normally,but when S_range = range(24),E_range = range(1,25),it reports an error.

David Torres Sanchez

2024年03月18日 11:19

Thanks for reporting this. We will track this in an internal ticket.

Cheers,
David

Jaromił Najman

2024年03月20日 09:54

It turned out to be an unhandled out-of-memory error. A fix will be available with version 11.0.2.

2024年03月20日 13:00

Thanks a lot!

Segmentation fault (core dumped)

コメント

お探しのものを見つけられませんでしたか？