Cross Platform Update

Added Cross Platform feature.

• You can now use different DB Providers through implementing the IRlmDbData interface
• Added built in implementation of SQL Server and PostGres

Added support for Gpu usage.

• Added built in Alea GPU

Projects

• Added RlmBasketballPrototype
• Added BasketballLineupOptimizer

Documentation

• Updated Developer Guide
• Updated Rlm Api Documentation

v1.0.7

RLM

• Data Persistence
  • Added RLM Data Persistence Feature toggle
  • Added Check on Data Persistence on Training Done
  • Added check to prevent calling on the DB Module when data persistence is off
• Transferred supported IDE to Visual Studio 2017 from 2015
• Improved Load Network
• Improved RNeuron query
• Bug Fixes:
  • Bug fix on Cancel Token disposed error
  • Bug fix on RLV query
  • Bug fix on RLM Training calling Training done method every after session
  • Bug fix on Load Network (Changed RNeuron and IVR lists to concurrent collections)
  • Bug fix on Input Momentum

Main Project:

• Added Challenger App (See Documentation)
• Added Retail PoC 2.0 with installer (See Documentation)

RetailPOC

• Added TensorFlow engine

RetailPOCSimple

• Commented out Debug Displays

RLV

• Added Web useable version
• Added Loading Screen and layout fixes
• Added new query

v1.0.6

• Added ENCOG engine to the RETAIL POC Application
• Added Ice Cream Application with Dynamic RLV

v1.0.5

• Added new RLM Learning Visualizer, which allows an insight towards the learning history of the RLM.

• Added various UI changes.

v1.0.4

• Fix an issue on the core wherein the benchmark stats were continually updating and stored big amounts of data in the background. Now, the benchmark tracking is set off by default and can be turned on programmatically for debugging purposes.
• Changed the Randomness used by RLM to a double value instead of integer to have a more accurate randomization chance especially when the randomness is set to a small range of values
• Initial optimizations for the RLM Memory which we will still need to continue to work on and update on the next release
• Updated the RetailPoC app and used optimized settings for the RLM.

v1.0.3

• Added new function to Reset the randomness counter of the RLM when doing multiple training sets
• Added code on the RLMDBEntities to enable us to no longer reference the EF on front-end apps when using the RLM

v1.0.2

• Minor fixes on the Data persistence progress
• Removed references for Encog and did some code cleanup for the example apps.
• Added a new example app, XOR.

v1.0

Fixed minor error on the sample apps