maximize_pairwise: Portfolio optimization using a pairwise algorithm

In rexmacey/TaxAwareAA: Tax Aware Asset Allocation

Description

This will attempt to maximize a supplied objective function within box constraints. The algorithm begins by finding all the asset classes with weights that can be increased and those that can be decreased. Using pairs of classes within the same account, it increases one and decreases the other finding the best pair to change. If a pair improves the objective value, the process is repeated. If no pair is found the maxdelta is halved and the process is repeated until the delta is <= 0.00001.

Usage

maximize_pairwise(x, LB, UB, objFun, maxdelta, ...)

Arguments

x	Initial set of weights. Must sum to 1 (or whatever sum is desired). Weights across accounts must be correct.
LB	Box minimum
UB	Box maximum
objFun	Objective function
maxdelta	Initial setting of delta.
...	Parameters to be passed to the objective function

Value

List with MaxObj the value of the objective; x the corresponding portfolio weights, and Iter the number of iterations

Examples

maximize_pairwise(x, LB, UB, objFun, maxdelta, ...)

rexmacey/TaxAwareAA documentation built on Dec. 3, 2019, 7:54 a.m.