RL-Ai-Player

• This is a exporation projection for classical reinforcement learning.

• The Algorithm we will be implemented are here

GYM starting

• to build all the images docker-compose bulid

• to run all the container run docker-compose up -d

• check for the tokens of jupyter notebook tokens with running docker-compose logs notebook to obtain ?token=SOME_HASH_STRING

• and visit localhost:8888?token=THE_TOKENS_THAT_SHOWN_IN_TERMINAL with your browser

Testing MDD with flink with docker compose

MapReduce Algorithm Steps

1. Map Function

   • Splitting
     • Splitting Step takes input DataSet from Source and divide into smaller Sub-DataSets.
   • Mapping
     • Mapping Step takes those smaller Sub-DataSets and perform required action or computation on each Sub-DataSet
   • This would finally pairs as <Key,Value> as output
   • Input: DataSets and Output List of <Key,Value>

2. Shuffle Function

   • Shuffling Function can explains as combine function
   • Merging combining all Key-Value pairs which have the same key and return <Key, List<Value>>.
   • Sorting based Key from aboved.

3. Reduce Function

   • Reduced Function Output generate List of <Key,Value>.