Hello everyone and happy new year,

I would like to use Gurobipy to find optimal portfolios by solving an integer program. Because the Gurobipy package or any similar Python package to optimise using the Gurobi engine are not available on the QuantConnect platform, I have purchased a Researcher seat, so that I can use Lean CLI on my computer. However, this also means that I need to subscribe to a data source, because data is only available for free when running backtests on the cloud. I have also watched a video introducing Lean CLI GUI by Colton (a very informative video), where he points out to the fact that a subscription to the QuantConnect Security Master database/dataset is necessary to run any backtest, because it is the one that contains the splits and dividends data.

Because I cannot afford a subscription to that data set (the project I'm working on is for an university seminar) but I need to use Gurobi for the optimisation (cvxpy is available for cloud backtests, but it is not a suitable alternative), I would like to know if there is:

  1. some way to perform cloud backtest using the Gurobipy package and Gurobi optimisation engine,
  2. or some way to perform local backtests using the Gurobipy package and Gurobi optimisation engine installed in my computer, without having to purchase such an expensive subscription (expensive for me).

 

My project consists in creating an algorithm to build portfolios that best track a given index, but with a limited number of assets. For example, construct portfolios of at most 50 assets that best track the returns of the S&P500. Therefore, I'm interested in using at least US equities data. Equities data from other regions or data of other asset classes would also be useful, but not strictly necessary. For example, in case we want to construct portfolios that best track the returns of the Norwegian Sovereign Wealth Fund.

Thank you very much for your ideas and help.

Author