BITA Optimisers

A fast family of flexible portfolio optimisers that handle complex problems, integrate easily and provide deep functionality.

The BITA Optimiser family consists of three different optimisers, each developed to provide specific approaches to portfolio construction. All are available as .dll's and are addressable from MatLab, S+, R, C#, .NET, Excel, Python, Perl etc. Common features are:

  • Fixed and flexible constraints
  • Full transaction cost handling (linear, piecewise, quadratic, long/short differential, own cost curve)
  • Full long short support and constraints
  • Custom & bespoke constraints

Quadratic Optimiser

Developed over the last 14 years the BITA Quadratic Optimiser is known for its flexibility, speed of processing and capacity to handle large complex problems. Designed to create portfolios and trades that are optimal in terms of risk, return and transaction costs, subject to a wide range of real-world trading and investment constraints, the most important applications of BITA Optimiser are in portfolio construction and back-testing. It also provides a library of portfolio risk analytics.

Robust

By applying Second Order Cone Programming Robust introduces one or more quadratic constraints on a portfolio. This enables specification of problems such as: degrees of uncertainty around alpha forecasts, consideration of historic forecast error, and ranked expected return. In addition optimisation with multiple risk models, or risk functions such as optimising for both tracking error and total risk simultaneously is possible. BITA Robust uses an interior point method and can be applied to sizable problems with reasonable speed.

GLO+

Our Gain/Loss Optimiser (GLO+) addresses higher moments going beyond Mean Variance. In pure form, GLO maximises expected gain and minimises expected loss subject to a given loss aversion utility. Extended to GLO+ it combines mean variance and gain/loss utilities in the objective function, enabling maximisation of return while minimising both variance and downside. Other functions can be combined in the utility.