使用Quantstrat进行投资组合优化

Question

I would like to backtest a portfolio of assets whose weights are mean-variance optimized.

I have seem examples here and here but all the examples are either dealing with fixed orderSize/tradeSize or flexible ordersize which is computed for 1 underlying asset only. However, mean-variance algorithm compute weights for a basket of assets (allocation for any assets should depends on its relation with other assets)

My understanding is that the osFUN parameter in the add.rule function returns order size for a particular symbol. Can it return a series of weights for all symbols? If so, how should 'osFUN' be structured? In a addition, will osFUN be called for every symbol on every trade date or be called only once for the portfolio overall?

Any other open-source software which can tackle this issue is welcome!

Thanks for helping!

Answer 1

我从QuantStrata博主那里得到了一个答案：“因为quantstrat不适用于此。Quantstrat旨在测试单个仪器上的信号过程。”