Execution Algorithms 102

Recent Changes

  • Changes to preserve the stats in the event of 'process restart'
  • Dividing the order in qleap-api itself
  • qleap-api tests
  • Single API server instance for multiple brokers

Planned Changes

  • We know it's really difficult to generate results for execution algorithms.
  • A function call to look up the execution cost for an order.

Planned Changes

  • Fix the incorrect results.

Planned Changes

  • Generate weekly models for 10 securities
  • Use model based execution algorithms

Future tasks/projects

  • Use common data/indicators in multiple models. This should help reduce the overhead due to indicator computation.
  • Try out reinforcement learning approach.
  • Add a processed text component.
Made with Slides.com