If the convex solver for subproblems accepts any additional keyword arguments, such as warm_start=True, then you can set them in the problem.solve() function, and they will be passed to the convex solver. The tau_max parameter upper bounds how large tau can get.The mu parameter sets the rate at which tau increases inside the CCP algorithm.Larger tau favors satisfying the constraints. The tau parameter trades off satisfying the constraints and minimizing the objective.The solver parameter specifies what solver to use to solve convex subproblems.The max_iter parameter sets the maximum number of iterations in the CCP algorithm.The ccp_times parameter specifies how many random initial points to run the algorithm from.Additional arguments can be used to specify the parameters.

problem.solve(method='dccp') applies the CCP heuristic, and returns the value of the cost function, the maximum value of the slack variables, and the value of each variable.Once a problem object has been constructed, the following solve method can be applied.

The components of the variable, the objective, and the constraints are constructed using standard CVXPY syntax. An option is to use random initialization for all variables by prob.solve(method=‘dccp’, random_start=TRUE), and by setting the parameter ccp_times you can specify the times that the CCP algorithm runs starting from random initial values for all variables each time.

What is concave programming