在cvar优化中写约束的有效方法

Question

基于此材料(see 5/6)，我实现了以下功能：

import pandas as pd

def optm_cvar(return_df, beta):
    import cvxpy
    from cvxpy import Variable
    from cvxpy import Minimize
    from cvxpy import Problem
    assert isinstance(return_df, pd.DataFrame)
    assert isinstance(beta, float)

    period, stock_num = return_df.shape

    weight = Variable(stock_num)
    value_at_risk = Variable()
    u = Variable(period)

    obj = Minimize(value_at_risk + (1 / ((1 - beta) * period)) * cvxpy.sum(u))

    constraint_l = list()
    constraint_l.append(weight >= 0)
    constraint_l.append(u >= 0)
    constraint_l.append(cvxpy.sum(weight) == 1)
    for i in range(period):
        constraint_l.append(u[i] + value_at_risk + cvxpy.sum(cvxpy.multiply(weight, return_df.iloc[i, :])) >= 0)

    problem = Problem(obj, constraint_l)

    #result = problem.solve()
    result = problem.solve(cvxpy.CVXOPT)
    #result = problem.solve(cvxpy.SCS)
    print(problem.status)
    print(result)
    return weight.value

此功能运行良好 但是我认为这个约束代码

for i in range(period):
    constraint_l.append(u[i] + value_at_risk + cvxpy.sum(cvxpy.multiply(weight, return_df.iloc[i, :])) >= 0)

效率不高，与cvxpy方式不匹配。 如何改善我的代码？