**PORTFOLIO CONSTRUCTION**

To optimize is "to make the best of." In practice, optimization involves more than diversifying firm-specific risks. Given all the conflicting objectives faced when building or rebalancing a portfolio - expected returns, the uncertainty attached to those, the relationships between the stocks and transaction costs - there are arguably too many complex trade-offs for the human mind to efficiently handle. For this reason mathematical tools are often employed to suggest possible trades, taking into account all of the above tradeoffs.

EM Applications' optimizer is designed to construct a portfolio which minimises risk whilst producing a certain level of expected performance. It does this by altering the proportions of each stock in the portfolio and determining the effect this has on the risk until it finds the best combination of stocks. The user may input a preference for each of the stocks which indicates the stock’s expected outperformance. It is also possible to specify the number of holdings desired within the final portfolio.

EMA’s optimizer uses some of the latest mathematical techniques to solve the typical problems faced by fund-managers. The core of a portfolio optimiser is based around two key algorithms – a quadratic-programming and an integer programming algorithm. Both of these routines have been applied to address the type of problems typical in asset management.