added mean variance builder

alphamind/portfolio/meanvariancebuilder.py

+# -*- coding: utf-8 -*-
+"""
+Created on 2017-6-27
+
+@author: cheng.li
+"""
+
+import numpy as np
+from typing import Union
+from typing import Tuple
+from cvxopt import matrix
+from cvxopt import solvers
+
+solvers.options['show_progress'] = False
+
+
+def mean_variance_builder(er: np.ndarray,
+                          cov: np.ndarray,
+                          bm: np.ndarray,
+                          lbound: Union[np.ndarray, float],
+                          ubound: Union[np.ndarray, float],
+                          risk_exposure: np.ndarray,
+                          risk_target: Tuple[np.ndarray, np.ndarray]):
+
+    lbound = lbound - bm
+    ubound = ubound - bm
+    transposed_risk_exposure = risk_exposure.T
+    risk_target = risk_target - transposed_risk_exposure @ bm
+
+    # set up problem for net position
+    n = len(er)
+
+    P = matrix(cov)
+    q = -matrix(er)
+
+    G1 = np.zeros((2*n, n))
+    h1 = np.zeros(2*n)
+
+    for i in range(n):
+        G1[i, i] = 1.
+        h1[i] = ubound[i]
+        G1[i+n, i] = -1.
+        h1[i+n] = -lbound[i]
+
+    m = len(transposed_risk_exposure)
+
+    G2 = np.concatenate([transposed_risk_exposure, -transposed_risk_exposure])
+    h2 = np.zeros(2*m)
+
+    for i in range(m):
+        h2[i] = risk_target[1][i]
+        h2[i+m] = -risk_target[0][i]
+
+    G = matrix(np.concatenate([G1, G2]))
+    h = matrix(np.concatenate([h1, h2]))
+
+    sol = solvers.qp(P, q, G, h)
+
+    return sol['status'], sol['dual objective'], np.array(sol['x']).flatten() + bm
+
+
+

alphamind/tests/portfolio/test_meanvariancebuild.py

+# -*- coding: utf-8 -*-
+"""
+Created on 2017-6-27
+
+@author: cheng.li
+"""
+
+import unittest
+import numpy as np
+from alphamind.portfolio.meanvariancebuilder import mean_variance_builder
+
+
+class TestMeanVarianceBuild(unittest.TestCase):
+
+    def test_mean_variance_builder(self):
+
+        er = np.random.randint(0, 10, size=3) / 10.
+        cov = np.array([[0.04, 0.01, 0.02],
+                        [0.01, 0.05, 0.03],
+                        [0.02, 0.03, 0.06]])
+        ids_var = np.diag(np.random.randint(2, 5, size=3) / 100.)
+        cov += ids_var
+
+        bm = np.array([0.3, 0.3, 0.4])
+        lbound = np.array([0., 0., 0.])
+        ubound = np.array([0.4, 0.4, 0.5])
+
+        risk_exposure = np.array([[1., 1., 1.],
+                                  [1., 0., 1.]]).T
+        risk_target = (np.array([bm.sum(), 0.3]), np.array([bm.sum(), 0.7]))
+
+        status, value, x = mean_variance_builder(er, cov, bm, lbound, ubound, risk_exposure, risk_target)
+
+        self.assertTrue(status == 'optimal')
+        self.assertAlmostEqual(x.sum(), bm.sum())
+        self.assertTrue(np.all(x <= ubound + 1.e-6))
+        self.assertTrue(np.all(x >= lbound) - 1.e-6)
+        self.assertTrue(np.all(x @ risk_exposure <= risk_target[1] + 1.e-6))
+        self.assertTrue(np.all(x @ risk_exposure >= risk_target[0] - 1.e-6))