​A selection of topics on IT and its application to finance. 
​Send me your comments, questions ​or suggestions by clicking

Portfolio optimisation.

The expected return of a portfolio of k assets is given by

µp = W’R


W’ is the transpose of the k x 1 series of asset weights which results in a 1 x k matrix and

R is the k x 1 matrix of asset returns.

Typically when optimising a portfolio of assets we will want to either

a) Maximise the return of the portfolio or

b) Minimise the risk of the portfolio (often we us
e the Sharpe ratio as a proxy for the portfolio risk)

So for each asset we need to calculate the average return and the Sharpe ratio given by



σp is the portfolio volatility                                                       
µp is the expected portfolio return and

rf is the risk free rate  e.g LIBOR , USA T-bill

In order to optimise (i.e maximise return/minimise Sharpe ratio) we make the assumption in the first instance that all the assets in the portfolio are equally weighted. We use these initial weights to calculate the portfolio return
µp and the portfoliovolatility σp.

Recall that
σp = √X’∑X where is the variance covariance matrix and X is the asset weight matrix and X' its transpose.

Now that we have all of this data we can use excel’s solver to do our optimisation by

either maximising
µp (or minimising the Sharpe ratio ) with the following constraints

     a)  Constraint 1  ->  SUM(weights)= 100%

     b)  Constraint 2  ->   Any individual weight > 0 (if we disallow short selling)

     c)  Constraint 3  ->  
σ <= smallest σ of any individual asset