Dictionary
A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z
Unconstrained Optimization
Definition
This is a statistical optimization technique that removes all constraints from the optimization process (with the exception that the weights must sum to 100%). If the optimization produces negative weights, this may indicate that the best action in a short position in thsse assets.
Using the term Unconstrained Optimization :
As part of their standard practice, the quant team always ran an unconstrained optimization first to see where the model wanted to go freely. After this excercise, constraints are added to create a more diverse portfolio in an effort to avoid too much concentration in one asset class.
Pay Special Attention To :
Unconstrained optimization rarely yields the correct allocation as it will often try to solve for the highest return at the expense of everything else. This usually puts the portfolio in very high concentrations of just one or two asset classes with the highest historical returns (and often the highest volatility).
Rate this definition:
Related terms
Statistics , Optimization Techniques , Sample Estimator

Dictionary