Abstract:
Some trading strategies are becoming more and more complicated
and utilize a large amount of data, which makes the backtesting
of these strategies very time consuming. This paper presents an efficient
implementation of the backtesting of such a trading strategy using a
parallel genetic algorithm (PGA) which is fine tuned based on thorough
analysis of the trading strategy. The reuse of intermediate results is very
important for such backtesting problems. Our implementation can perform
the back testing within a reasonable time range so that the tested
trading strategy can be properly deployed in time.
