def test_abs_expr_to_grb_expr(self):
"""
min |x + 1| s.t. x <= -4
"""
aff = AffExpr(np.ones((1,1)), np.ones((1,1)))
abs_expr = AbsExpr(aff)
aff = AffExpr(np.ones((1,1)), np.zeros((1,1)))
comp = LEqExpr(aff, np.array([[-4]]))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
model.update()
abs_grb_expr, abs_grb_cnt = prob._abs_expr_to_grb_expr(abs_expr, var)
model.update()
model.setObjective(abs_grb_expr[0,0])
bexpr = BoundExpr(comp, var)
prob.add_cnt_expr(bexpr)
model.optimize()
var.update()
self.assertTrue(np.allclose(var.get_value(), np.array([[-4]])))
# makes assumption about the construction of the Gurobi variable, needs
# to be changed TODO
pos = abs_grb_expr[0,0].getVar(0).X
neg = abs_grb_expr[0,0].getVar(1).X
self.assertTrue(np.allclose(pos, 0.0))
self.assertTrue(np.allclose(neg, 3.0))
def test_hinge_expr_to_grb_expr2(self):
"""
min max(0, x+1) st. x == 1
"""
aff = AffExpr(np.ones((1, 1)), np.ones((1, 1)))
hinge = HingeExpr(aff)
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
model.update()
hinge_grb_expr, hinge_grb_cnt = prob._hinge_expr_to_grb_expr(
hinge, var)
model.update()
obj = hinge_grb_expr[0, 0]
model.setObjective(obj)
aff = AffExpr(np.ones((1, 1)), np.zeros((1, 1)))
comp = EqExpr(aff, np.array([[1.0]]))
bound_expr = BoundExpr(comp, var)
prob.add_cnt_expr(bound_expr)
model.optimize()
var.update()
self.assertTrue(np.allclose(var.get_value(), np.array([[1.0]])))
self.assertTrue(np.allclose(obj.X, 2.0))
def test_hinge_expr_to_grb_expr2(self):
"""
min max(0, x+1) st. x == 1
"""
aff = AffExpr(np.ones((1,1)), np.ones((1,1)))
hinge = HingeExpr(aff)
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
model.update()
hinge_grb_expr, hinge_grb_cnt = prob._hinge_expr_to_grb_expr(hinge, var)
model.update()
obj = hinge_grb_expr[0,0]
model.setObjective(obj)
aff = AffExpr(np.ones((1,1)), np.zeros((1,1)))
comp = EqExpr(aff, np.array([[1.0]]))
bound_expr = BoundExpr(comp, var)
prob.add_cnt_expr(bound_expr)
model.optimize()
var.update()
self.assertTrue(np.allclose(var.get_value(), np.array([[1.0]])))
self.assertTrue(np.allclose(obj.X, 2.0))
def __init__(self,
num_nodes,
num_bits,
node_capacity,
conservation_factor,
confidence_factor,
init_num_nodes,
alpha,
beta,
gamma,
delta_in,
delta_out,
plot_degree_dist=False):
self.num_nodes = num_nodes
self.num_bits = num_bits
self.prob = Prob(num_bits, conservation_factor, confidence_factor)
self.scale_free_graph = generate_scale_free(num_nodes, init_num_nodes,
alpha, beta, gamma,
delta_in, delta_out)
self.nodes_memory = [
MemoryQueue(node_capacity) for _ in range(num_nodes)
]
if plot_degree_dist:
plot_degree_distribution(
self.scale_free_graph.compute_outdegree_distribution(),
int(0.08 * num_nodes), 'Outdegree Distribution')
plot_degree_distribution(
self.scale_free_graph.compute_indegree_distribution(),
int(0.08 * num_nodes), 'Indegree Distribution')
def test_abs_expr_to_grb_expr(self):
"""
min |x + 1| s.t. x <= -4
"""
aff = AffExpr(np.ones((1, 1)), np.ones((1, 1)))
abs_expr = AbsExpr(aff)
aff = AffExpr(np.ones((1, 1)), np.zeros((1, 1)))
comp = LEqExpr(aff, np.array([[-4]]))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
model.update()
abs_grb_expr, abs_grb_cnt = prob._abs_expr_to_grb_expr(abs_expr, var)
model.update()
model.setObjective(abs_grb_expr[0, 0])
bexpr = BoundExpr(comp, var)
prob.add_cnt_expr(bexpr)
model.optimize()
var.update()
self.assertTrue(np.allclose(var.get_value(), np.array([[-4]])))
# makes assumption about the construction of the Gurobi variable, needs
# to be changed TODO
pos = abs_grb_expr[0, 0].getVar(0).X
neg = abs_grb_expr[0, 0].getVar(1).X
self.assertTrue(np.allclose(pos, 0.0))
self.assertTrue(np.allclose(neg, 3.0))
def test_expr_to_grb_expr_w_expr(self):
model = grb.Model()
prob = Prob(model)
expr = Expr(f)
bexpr = BoundExpr(f, None)
with self.assertRaises(Exception) as e:
prob._expr_to_grb_expr(bexpr)
self.assertTrue("Expression cannot be converted" in e.exception.message)
def test_expr_to_grb_expr_w_expr(self):
model = grb.Model()
prob = Prob(model)
expr = Expr(f)
bexpr = BoundExpr(f, None)
with self.assertRaises(Exception) as e:
prob._expr_to_grb_expr(bexpr)
self.assertTrue(
"Expression cannot be converted" in e.exception.message)
def test_expr_to_grb_expr_w_comp_expr(self):
model = grb.Model()
prob = Prob(model)
aff = AffExpr(-2 * np.ones((1, 1)), np.zeros((1, 1)))
val = np.zeros((1, 1))
cexpr = CompExpr(aff, val)
bexpr = BoundExpr(cexpr, None)
with self.assertRaises(Exception) as e:
prob._expr_to_grb_expr(bexpr)
self.assertTrue("Comparison" in e.exception.message)
def test_expr_to_grb_expr_w_comp_expr(self):
model = grb.Model()
prob = Prob(model)
aff = AffExpr(-2*np.ones((1,1)), np.zeros((1,1)))
val = np.zeros((1,1))
cexpr = CompExpr(aff, val)
bexpr = BoundExpr(cexpr, None)
with self.assertRaises(Exception) as e:
prob._expr_to_grb_expr(bexpr)
self.assertTrue("Comparison" in e.exception.message)
def test_get_max_cnt_violation_leq_cnts(self):
model = grb.Model()
prob = Prob(model)
dummy_var = Variable(np.zeros((1, 1)), np.zeros((1, 1)))
f = lambda x: np.array([[1, 3]])
f_expr = Expr(f)
leq_expr = LEqExpr(f_expr, np.array([[1, 1]]))
bexpr = BoundExpr(leq_expr, dummy_var)
prob.add_cnt_expr(bexpr)
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 2.0))
f_expr = Expr(f)
f_expr.f = lambda x: np.array([[2, 1]])
leq_expr.expr = f_expr
leq_expr.val = np.array([[1, 1]])
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 1.0))
f_expr = Expr(f)
f_expr.f = lambda x: np.array([[2, -2]])
leq_expr.expr = f_expr
leq_expr.val = np.array([[1, 1]])
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 1.0))
f_expr = Expr(f)
f_expr.f = lambda x: np.array([[2, -2]])
leq_expr.expr = f_expr
leq_expr.val = np.array([[2, -2]])
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 0.0))
f_expr = Expr(f)
f_expr.f = lambda x: np.array([[2, 0]])
leq_expr.expr = f_expr
leq_expr.val = np.array([[2, -2]])
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 2.0))
def test_callback(self):
x = {}
def test():
x[1] = 2
callback = test
model = grb.Model()
prob = Prob(model, callback)
prob.find_closest_feasible_point()
self.assertTrue(1 in x)
x[1] = 3
prob.optimize()
self.assertTrue(1 in x)
def test_add_obj_expr_nonquad(self):
expr = Expr(f)
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
bexpr = BoundExpr(expr, var)
prob.add_obj_expr(bexpr)
self.assertTrue(bexpr not in prob._quad_obj_exprs)
self.assertTrue(bexpr in prob._nonquad_obj_exprs)
self.assertTrue(var in prob._vars)
def test_get_max_cnt_violation_leq_cnts(self):
model = grb.Model()
prob = Prob(model)
dummy_var = Variable(np.zeros((1,1)), np.zeros((1,1)))
f = lambda x: np.array([[1,3]])
f_expr = Expr(f)
leq_expr = LEqExpr(f_expr, np.array([[1,1]]))
bexpr = BoundExpr(leq_expr, dummy_var)
prob.add_cnt_expr(bexpr)
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 2.0))
f_expr = Expr(f)
f_expr.f = lambda x: np.array([[2,1]])
leq_expr.expr = f_expr
leq_expr.val = np.array([[1,1]])
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 1.0))
f_expr = Expr(f)
f_expr.f = lambda x: np.array([[2,-2]])
leq_expr.expr = f_expr
leq_expr.val = np.array([[1,1]])
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 1.0))
f_expr = Expr(f)
f_expr.f = lambda x: np.array([[2,-2]])
leq_expr.expr = f_expr
leq_expr.val = np.array([[2, -2]])
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 0.0))
f_expr = Expr(f)
f_expr.f = lambda x: np.array([[2, 0]])
leq_expr.expr = f_expr
leq_expr.val = np.array([[2, -2]])
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 2.0))
def test_optimize_multidim_quad_obj(self):
Q = np.array([[2, 0], [0, 0]])
A = np.array([[-4, 0]])
quad = QuadExpr(Q, A, np.zeros((1, 1)))
model = grb.Model()
prob = Prob(model)
grb_var1 = model.addVar(lb=-1 * GRB.INFINITY,
ub=GRB.INFINITY,
name='x1')
grb_var2 = model.addVar(lb=-1 * GRB.INFINITY,
ub=GRB.INFINITY,
name='x2')
grb_vars = np.array([[grb_var1], [grb_var2]])
var = Variable(grb_vars)
bexpr_quad = BoundExpr(quad, var)
prob.add_obj_expr(bexpr_quad)
self.assertTrue(bexpr_quad in prob._quad_obj_exprs)
self.assertTrue(var in prob._vars)
prob.update_obj(penalty_coeff=0)
prob.optimize()
var_value = var.get_value()
value = np.zeros((2, 1))
value[0, 0] = 2
self.assertTrue(np.allclose(var_value, value))
def test_add_obj_expr_nonquad(self):
expr = Expr(f)
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
bexpr = BoundExpr(expr, var)
prob.add_obj_expr(bexpr)
self.assertTrue(bexpr not in prob._quad_obj_exprs)
self.assertTrue(bexpr in prob._nonquad_obj_exprs)
self.assertTrue(var in prob._vars)
def test_add_cnt_expr_eq_aff(self):
aff = AffExpr(np.ones((1,1)), np.zeros((1,1)))
comp = EqExpr(aff, np.array([[2]]))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
model.update()
bexpr = BoundExpr(comp, var)
prob.add_cnt_expr(bexpr)
prob.optimize()
self.assertTrue(np.allclose(var.get_value(), np.array([[2]])))
def test_add_obj_expr_quad(self):
quad = QuadExpr(2*np.eye(1), -2*np.ones((1,1)), np.zeros((1,1)))
aff = AffExpr(-2*np.ones((1,1)), np.zeros((1,1)))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
bexpr_quad = BoundExpr(quad, var)
bexpr_aff = BoundExpr(aff, var)
prob.add_obj_expr(bexpr_quad)
prob.add_obj_expr(bexpr_aff)
self.assertTrue(bexpr_aff in prob._quad_obj_exprs)
self.assertTrue(bexpr_quad in prob._quad_obj_exprs)
self.assertTrue(var in prob._vars)
def test_find_closest_feasible_point_eq_cnts(self):
model = grb.Model()
prob = Prob(model)
grb_var1 = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x1')
grb_var2 = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x2')
grb_vars = np.array([[grb_var1],[grb_var2]])
var = Variable(grb_vars, np.zeros((2,1)))
model.update()
val = np.array([[5.0],[-10.0]])
aff_expr = AffExpr(np.eye(2), np.zeros((2,1)))
eq_expr = EqExpr(aff_expr, val)
bexpr = BoundExpr(eq_expr, var)
prob.add_cnt_expr(bexpr)
prob.find_closest_feasible_point()
self.assertTrue(np.allclose(var.get_value(), val))
def test_convexify_leq(self):
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
model.update()
grb_cnt = model.addConstr(grb_var, GRB.EQUAL, 0)
model.optimize()
var.update()
e = Expr(f)
eq = LEqExpr(e, np.array([[4]]))
bexpr = BoundExpr(eq, var)
prob.add_cnt_expr(bexpr)
prob.convexify()
self.assertTrue(len(prob._penalty_exprs) == 1)
self.assertTrue(isinstance(prob._penalty_exprs[0].expr, HingeExpr))
def test_optimize_just_quad_obj(self):
quad = QuadExpr(2 * np.eye(1), -2 * np.ones((1, 1)), np.zeros((1, 1)))
aff = AffExpr(-2 * np.ones((1, 1)), np.zeros((1, 1)))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
bexpr_quad = BoundExpr(quad, var)
bexpr_aff = BoundExpr(aff, var)
prob.add_obj_expr(bexpr_quad)
prob.add_obj_expr(bexpr_aff)
self.assertTrue(bexpr_aff in prob._quad_obj_exprs)
self.assertTrue(bexpr_quad in prob._quad_obj_exprs)
self.assertTrue(var in prob._vars)
prob.update_obj(penalty_coeff=0)
prob.optimize()
self.assertTrue(np.allclose(var.get_value(), np.array([[2.0]])))
def test_find_closest_feasible_point_leq_cnts(self):
cnt_vals = [np.ones((2,1)), np.array([[-1.0],[1.0]]), \
np.array([[-1.0],[-1.0]])]
true_var_vals = [np.zeros((2,1)), np.array([[-1.0],[0.0]]), \
-1*np.ones((2,1))]
for true_var_val, cnt_val in zip(true_var_vals, cnt_vals):
model = grb.Model()
prob = Prob(model)
grb_var1 = model.addVar(lb=-1 * GRB.INFINITY,
ub=GRB.INFINITY,
name='x1')
grb_var2 = model.addVar(lb=-1 * GRB.INFINITY,
ub=GRB.INFINITY,
name='x2')
grb_vars = np.array([[grb_var1], [grb_var2]])
var = Variable(grb_vars, np.zeros((2, 1)))
model.update()
aff_expr = AffExpr(np.eye(2), np.zeros((2, 1)))
leq_expr = LEqExpr(aff_expr, cnt_val)
bexpr = BoundExpr(leq_expr, var)
prob.add_cnt_expr(bexpr)
prob.find_closest_feasible_point()
self.assertTrue(np.allclose(var.get_value(), true_var_val))
def test_prob(ut,
x0,
x_true,
f=zerofunc,
g=neginffunc,
h=zerofunc,
Q=np.zeros((N, N)),
q=np.zeros((1, N)),
A_ineq=np.zeros((1, N)),
b_ineq=np.zeros((1, 1)),
A_eq=np.zeros((1, 1)),
b_eq=np.zeros((1, 1))):
if not np.allclose(A_eq, np.zeros((1,1)))\
or not np.allclose(b_eq, np.zeros((1,1))):
raise NotImplementedError
model = grb.Model()
prob = Prob(model)
grb_var1 = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x1')
grb_var2 = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x2')
grb_vars = np.array([[grb_var1], [grb_var2]])
var = Variable(grb_vars, value=x0)
model.update()
quad_obj = BoundExpr(QuadExpr(Q, q, np.zeros((1, 1))), var)
prob.add_obj_expr(quad_obj)
nonquad_obj = BoundExpr(Expr(f), var)
prob.add_obj_expr(nonquad_obj)
cnts = []
lin_ineq = LEqExpr(AffExpr(A_ineq, -b_ineq), np.zeros(b_ineq.shape))
lin_ineq = BoundExpr(lin_ineq, var)
cnts.append(lin_ineq)
nonlin_ineq = LEqExpr(Expr(g), np.zeros(g(np.zeros((2, 1))).shape))
nonlin_ineq = BoundExpr(nonlin_ineq, var)
cnts.append(nonlin_ineq)
nonlin_eq = EqExpr(Expr(h), np.zeros(g(np.zeros((2, 1))).shape))
nonlin_eq = BoundExpr(nonlin_eq, var)
cnts.append(nonlin_eq)
for cnt in cnts:
prob.add_cnt_expr(cnt)
solv.solve(prob, method='penalty_sqp')
x_sol = var.get_value()
ut.assertTrue(np.allclose(x_sol, x_true, atol=1e-4))
def test_find_closest_feasible_point_leq_cnts(self):
cnt_vals = [np.ones((2,1)), np.array([[-1.0],[1.0]]), \
np.array([[-1.0],[-1.0]])]
true_var_vals = [np.zeros((2,1)), np.array([[-1.0],[0.0]]), \
-1*np.ones((2,1))]
for true_var_val, cnt_val in zip(true_var_vals, cnt_vals):
model = grb.Model()
prob = Prob(model)
grb_var1 = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x1')
grb_var2 = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x2')
grb_vars = np.array([[grb_var1],[grb_var2]])
var = Variable(grb_vars, np.zeros((2,1)))
model.update()
aff_expr = AffExpr(np.eye(2), np.zeros((2,1)))
leq_expr = LEqExpr(aff_expr, cnt_val)
bexpr = BoundExpr(leq_expr, var)
prob.add_cnt_expr(bexpr)
prob.find_closest_feasible_point()
self.assertTrue(np.allclose(var.get_value(), true_var_val))
def test_optimize_just_quad_obj(self):
quad = QuadExpr(2*np.eye(1), -2*np.ones((1,1)), np.zeros((1,1)))
aff = AffExpr(-2*np.ones((1,1)), np.zeros((1,1)))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
bexpr_quad = BoundExpr(quad, var)
bexpr_aff = BoundExpr(aff, var)
prob.add_obj_expr(bexpr_quad)
prob.add_obj_expr(bexpr_aff)
self.assertTrue(bexpr_aff in prob._quad_obj_exprs)
self.assertTrue(bexpr_quad in prob._quad_obj_exprs)
self.assertTrue(var in prob._vars)
prob.update_obj(penalty_coeff=0)
prob.optimize()
self.assertTrue(np.allclose(var.get_value(), np.array([[2.0]])))
def test_add_cnt_leq_aff(self):
quad = QuadExpr(2 * np.eye(1), -2 * np.ones((1, 1)), np.zeros((1, 1)))
aff = AffExpr(np.ones((1, 1)), np.zeros((1, 1)))
comp = LEqExpr(aff, np.array([[-4]]))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
model.update()
bexpr_quad = BoundExpr(quad, var)
prob.add_obj_expr(bexpr_quad)
bexpr = BoundExpr(comp, var)
prob.add_cnt_expr(bexpr)
prob.optimize()
self.assertTrue(np.allclose(var.get_value(), np.array([[-4]])))
def test_get_max_cnt_violation_mult_cnts(self):
model = grb.Model()
prob = Prob(model)
dummy_var = Variable(np.zeros((1, 1)), np.zeros((1, 1)))
f1 = lambda x: np.array([[1, 3]])
f2 = lambda x: np.array([[0, 0]])
f1_expr = Expr(f1)
leq_expr = LEqExpr(f1_expr, np.array([[1, 1]]))
bexpr = BoundExpr(leq_expr, dummy_var)
prob.add_cnt_expr(bexpr)
f2_expr = Expr(f2)
eq_expr = EqExpr(f2_expr, np.array([[1, 1]]))
bexpr = BoundExpr(eq_expr, dummy_var)
prob.add_cnt_expr(bexpr)
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 2.0))
def test_callback(self):
x = {}
def test():
x[1] = 2
callback = test
model = grb.Model()
prob = Prob(model, callback)
prob.find_closest_feasible_point()
self.assertTrue(1 in x)
x[1] = 3
prob.optimize()
self.assertTrue(1 in x)
def test_pos_grb_var_manager(self):
model = grb.Model()
prob = Prob(model)
init_num = 1
inc_num = 10
shape = (2, 7)
pos_grb_manager = PosGRBVarManager(model,
init_num=init_num,
inc_num=inc_num)
pgm = pos_grb_manager
PosGRBVarManager.INC_NUM = 2
self.assertTrue(len(pgm._grb_vars) == init_num)
var = pgm.next()
test_grb_var_pos(self, var)
self.assertTrue(len(pgm._grb_vars) == inc_num + init_num)
a = pgm.get_array(shape)
for x in a.flatten():
test_grb_var_pos(self, x)
self.assertTrue(a.shape == shape)
self.assertTrue(len(pgm._grb_vars) == init_num + 2 * inc_num)
def test_add_obj_expr_quad(self):
quad = QuadExpr(2 * np.eye(1), -2 * np.ones((1, 1)), np.zeros((1, 1)))
aff = AffExpr(-2 * np.ones((1, 1)), np.zeros((1, 1)))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
bexpr_quad = BoundExpr(quad, var)
bexpr_aff = BoundExpr(aff, var)
prob.add_obj_expr(bexpr_quad)
prob.add_obj_expr(bexpr_aff)
self.assertTrue(bexpr_aff in prob._quad_obj_exprs)
self.assertTrue(bexpr_quad in prob._quad_obj_exprs)
self.assertTrue(var in prob._vars)
def test_prob(ut, x0, x_true, f=zerofunc, g=neginffunc, h=zerofunc,
Q=np.zeros((N,N)), q=np.zeros((1,N)), A_ineq=np.zeros((1,N)),
b_ineq=np.zeros((1,1)), A_eq=np.zeros((1,1)), b_eq=np.zeros((1,1))):
if not np.allclose(A_eq, np.zeros((1,1)))\
or not np.allclose(b_eq, np.zeros((1,1))):
raise NotImplementedError
model = grb.Model()
prob = Prob(model)
grb_var1 = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x1')
grb_var2 = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x2')
grb_vars = np.array([[grb_var1], [grb_var2]])
var = Variable(grb_vars, value=x0)
model.update()
quad_obj = BoundExpr(QuadExpr(Q, q, np.zeros((1,1))), var)
prob.add_obj_expr(quad_obj)
nonquad_obj = BoundExpr(Expr(f), var)
prob.add_obj_expr(nonquad_obj)
cnts = []
lin_ineq = LEqExpr(AffExpr(A_ineq, -b_ineq), np.zeros(b_ineq.shape))
lin_ineq = BoundExpr(lin_ineq, var)
cnts.append(lin_ineq)
nonlin_ineq = LEqExpr(Expr(g), np.zeros(g(np.zeros((2,1))).shape))
nonlin_ineq = BoundExpr(nonlin_ineq, var)
cnts.append(nonlin_ineq)
nonlin_eq = EqExpr(Expr(h), np.zeros(g(np.zeros((2,1))).shape))
nonlin_eq = BoundExpr(nonlin_eq, var)
cnts.append(nonlin_eq)
for cnt in cnts:
prob.add_cnt_expr(cnt)
solv.solve(prob, method='penalty_sqp')
x_sol = var.get_value()
ut.assertTrue(np.allclose(x_sol, x_true, atol=1e-4))
def test_get_max_cnt_violation_mult_cnts(self):
model = grb.Model()
prob = Prob(model)
dummy_var = Variable(np.zeros((1,1)), np.zeros((1,1)))
f1 = lambda x: np.array([[1,3]])
f2 = lambda x: np.array([[0,0]])
f1_expr = Expr(f1)
leq_expr = LEqExpr(f1_expr, np.array([[1,1]]))
bexpr = BoundExpr(leq_expr, dummy_var)
prob.add_cnt_expr(bexpr)
f2_expr = Expr(f2)
eq_expr = EqExpr(f2_expr, np.array([[1,1]]))
bexpr = BoundExpr(eq_expr, dummy_var)
prob.add_cnt_expr(bexpr)
self.assertTrue(np.allclose(prob.get_max_cnt_violation(), 2.0))
def test_optimize_multidim_quad_obj(self):
Q = np.array([[2,0], [0,0]])
A = np.array([[-4, 0]])
quad = QuadExpr(Q, A, np.zeros((1,1)))
model = grb.Model()
prob = Prob(model)
grb_var1 = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x1')
grb_var2 = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x2')
grb_vars = np.array([[grb_var1], [grb_var2]])
var = Variable(grb_vars)
bexpr_quad = BoundExpr(quad, var)
prob.add_obj_expr(bexpr_quad)
self.assertTrue(bexpr_quad in prob._quad_obj_exprs)
self.assertTrue(var in prob._vars)
prob.update_obj(penalty_coeff=0)
prob.optimize()
var_value = var.get_value()
value = np.zeros((2,1))
value[0,0] = 2
self.assertTrue(np.allclose(var_value, value))
def test_convexify_leq(self):
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
model.update()
grb_cnt = model.addConstr(grb_var, GRB.EQUAL, 0)
model.optimize()
var.update()
e = Expr(f)
eq = LEqExpr(e, np.array([[4]]))
bexpr = BoundExpr(eq, var)
prob.add_cnt_expr(bexpr)
prob.convexify()
self.assertTrue(len(prob._penalty_exprs) == 1)
self.assertTrue(isinstance(prob._penalty_exprs[0].expr, HingeExpr))
def test_find_closest_feasible_point_eq_cnts(self):
model = grb.Model()
prob = Prob(model)
grb_var1 = model.addVar(lb=-1 * GRB.INFINITY,
ub=GRB.INFINITY,
name='x1')
grb_var2 = model.addVar(lb=-1 * GRB.INFINITY,
ub=GRB.INFINITY,
name='x2')
grb_vars = np.array([[grb_var1], [grb_var2]])
var = Variable(grb_vars, np.zeros((2, 1)))
model.update()
val = np.array([[5.0], [-10.0]])
aff_expr = AffExpr(np.eye(2), np.zeros((2, 1)))
eq_expr = EqExpr(aff_expr, val)
bexpr = BoundExpr(eq_expr, var)
prob.add_cnt_expr(bexpr)
prob.find_closest_feasible_point()
self.assertTrue(np.allclose(var.get_value(), val))
def test_get_approx_value_lin_constr(self):
"""
min x^2 st. x == 4
when convexified,
min x^2 + penalty_coeff*|x-4|
when penalty_coeff == 1, solution is x = 0.5 and the value is 3.75
(according to Wolfram Alpha)
when penalty_coeff == 2, solution is x = 1.0 and the value is 7.0
(according to Wolfram Alpha)
"""
quad = QuadExpr(2*np.eye(1), np.zeros((1,1)), np.zeros((1,1)))
e = Expr(f)
eq = EqExpr(e, np.array([[4]]))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
model.update()
obj = BoundExpr(quad, var)
prob.add_obj_expr(obj)
bexpr = BoundExpr(eq, var)
prob.add_cnt_expr(bexpr)
prob.optimize() # needed to set an initial value
prob.convexify()
prob.update_obj(penalty_coeff=1.0)
prob.optimize()
self.assertTrue(np.allclose(var.get_value(), np.array([[0.5]])))
self.assertTrue(np.allclose(prob.get_approx_value(1.0), np.array([[3.75]])))
prob.update_obj(penalty_coeff=2.0)
prob.optimize()
self.assertTrue(np.allclose(var.get_value(), np.array([[1.0]])))
self.assertTrue(np.allclose(prob.get_approx_value(2.0), np.array([[7]])))
def test_get_value_and_get_approx_value_nonlin_constr(self):
"""
min x^2 -2x + 1 st. x^2 == 4
when convexified at x = 1,
min x^2 -2x + 1 + penalty_coeff*|2x-5|
when penalty_coeff == 0.5, solution is x = 1.5 and the value is 1.25
(according to Wolfram Alpha)
approx value should be 1.25
value should be 1.125
"""
quad = QuadExpr(2*np.eye(1), -2*np.ones((1,1)), np.ones((1,1)))
quad_cnt = QuadExpr(2*np.eye(1), np.zeros((1,1)), np.zeros((1,1)))
eq = EqExpr(quad_cnt, np.array([[4]]))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars, np.array([[1.0]]))
model.update()
obj = BoundExpr(quad, var)
prob.add_obj_expr(obj)
bexpr = BoundExpr(eq, var)
prob.add_cnt_expr(bexpr)
prob.convexify()
prob.update_obj(penalty_coeff=0.5)
prob.optimize()
self.assertTrue(np.allclose(var.get_value(), np.array([[1.5]])))
self.assertTrue(np.allclose(prob.get_approx_value(0.5), np.array([[1.25]])))
self.assertTrue(np.allclose(prob.get_value(0.5), np.array([[1.125]])))
class RumourSpreadModel:
def __init__(self,
num_nodes,
num_bits,
node_capacity,
conservation_factor,
confidence_factor,
init_num_nodes,
alpha,
beta,
gamma,
delta_in,
delta_out,
plot_degree_dist=False):
self.num_nodes = num_nodes
self.num_bits = num_bits
self.prob = Prob(num_bits, conservation_factor, confidence_factor)
self.scale_free_graph = generate_scale_free(num_nodes, init_num_nodes,
alpha, beta, gamma,
delta_in, delta_out)
self.nodes_memory = [
MemoryQueue(node_capacity) for _ in range(num_nodes)
]
if plot_degree_dist:
plot_degree_distribution(
self.scale_free_graph.compute_outdegree_distribution(),
int(0.08 * num_nodes), 'Outdegree Distribution')
plot_degree_distribution(
self.scale_free_graph.compute_indegree_distribution(),
int(0.08 * num_nodes), 'Indegree Distribution')
def get_graph(self):
return self.scale_free_graph
def get_nodes_memory(self):
return self.nodes_memory
def distort_info(self, x):
bit_pos = randint(0, self.num_bits - 1)
return (x ^ (1 << bit_pos))
def get_distorted_info(self):
distorted_info = []
for node_memory in self.nodes_memory:
if node_memory.is_empty():
distorted_info.append(None)
continue
most_freq_info = node_memory.get_most_freq_elem()
distortion_prob = self.prob.compute_distortion_prob(
node_memory.compute_entropy())
if random() <= distortion_prob:
dist_info = self.distort_info(most_freq_info)
node_memory.distort_in_memory(most_freq_info, dist_info)
distorted_info.append(dist_info)
else:
distorted_info.append(most_freq_info)
return distorted_info
def update_node_memory(self, buffer_info):
nodes_outdegree = self.scale_free_graph.compute_outdegrees()
max_nbr_outdegree = [0] * self.num_nodes
min_nbr_outdegree = [self.num_nodes] * self.num_nodes
for node in range(self.num_nodes):
nbr_list = self.scale_free_graph.adj_list[node]
degree = nodes_outdegree[node]
for nbr in nbr_list:
max_nbr_outdegree[nbr] = max(max_nbr_outdegree[nbr], degree)
min_nbr_outdegree[nbr] = min(min_nbr_outdegree[nbr], degree)
for node in range(self.num_nodes):
if buffer_info[node] is None:
continue
nbr_list = self.scale_free_graph.adj_list[node]
degree = nodes_outdegree[node]
for nbr in nbr_list:
acceptance_prob = self.prob.compute_acceptance_prob(
degree, max_nbr_outdegree[nbr], min_nbr_outdegree[nbr])
if random() <= acceptance_prob:
self.nodes_memory[nbr].insert(buffer_info[node])
def inject_info(self, info_propagators):
for node, info in info_propagators.items():
self.nodes_memory[node].insert(info)
def simulate_step(self):
distorted_info = self.get_distorted_info()
self.update_node_memory(distorted_info)
def simulate(self, info_propagators, time_steps, callback_list):
self.inject_info(info_propagators)
for t in range(time_steps):
self.simulate_step()
for callback in callback_list:
callback.call_after_step(self, t)
return [callback.get_result() for callback in callback_list]
def test_get_value_and_get_approx_value_nonlin_constr(self):
"""
min x^2 -2x + 1 st. x^2 == 4
when convexified at x = 1,
min x^2 -2x + 1 + penalty_coeff*|2x-5|
when penalty_coeff == 0.5, solution is x = 1.5 and the value is 1.25
(according to Wolfram Alpha)
approx value should be 1.25
value should be 1.125
"""
quad = QuadExpr(2 * np.eye(1), -2 * np.ones((1, 1)), np.ones((1, 1)))
quad_cnt = QuadExpr(2 * np.eye(1), np.zeros((1, 1)), np.zeros((1, 1)))
eq = EqExpr(quad_cnt, np.array([[4]]))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars, np.array([[1.0]]))
model.update()
obj = BoundExpr(quad, var)
prob.add_obj_expr(obj)
bexpr = BoundExpr(eq, var)
prob.add_cnt_expr(bexpr)
prob.convexify()
prob.update_obj(penalty_coeff=0.5)
prob.optimize()
self.assertTrue(np.allclose(var.get_value(), np.array([[1.5]])))
self.assertTrue(
np.allclose(prob.get_approx_value(0.5), np.array([[1.25]])))
self.assertTrue(np.allclose(prob.get_value(0.5), np.array([[1.125]])))
def test_get_approx_value_lin_constr(self):
"""
min x^2 st. x == 4
when convexified,
min x^2 + penalty_coeff*|x-4|
when penalty_coeff == 1, solution is x = 0.5 and the value is 3.75
(according to Wolfram Alpha)
when penalty_coeff == 2, solution is x = 1.0 and the value is 7.0
(according to Wolfram Alpha)
"""
quad = QuadExpr(2 * np.eye(1), np.zeros((1, 1)), np.zeros((1, 1)))
e = Expr(f)
eq = EqExpr(e, np.array([[4]]))
model = grb.Model()
prob = Prob(model)
grb_var = model.addVar(lb=-1 * GRB.INFINITY, ub=GRB.INFINITY, name='x')
grb_vars = np.array([[grb_var]])
var = Variable(grb_vars)
model.update()
obj = BoundExpr(quad, var)
prob.add_obj_expr(obj)
bexpr = BoundExpr(eq, var)
prob.add_cnt_expr(bexpr)
prob.optimize() # needed to set an initial value
prob.convexify()
prob.update_obj(penalty_coeff=1.0)
prob.optimize()
self.assertTrue(np.allclose(var.get_value(), np.array([[0.5]])))
self.assertTrue(
np.allclose(prob.get_approx_value(1.0), np.array([[3.75]])))
prob.update_obj(penalty_coeff=2.0)
prob.optimize()
self.assertTrue(np.allclose(var.get_value(), np.array([[1.0]])))
self.assertTrue(
np.allclose(prob.get_approx_value(2.0), np.array([[7]])))
