Methods and apparatus to adaptively optimize feed blend and medicinal selection using machine learning to optimize animal milk production and health

An apparatus, comprising:

• a memory;
• and
• a processor operatively coupled to the memory, the processor configured to:
• receive a first value associated with a property of an environment in which an agent is deployed at a first state from a set of states defined within the environment, the agent configured to interact with the environment by performing at least one action selected from a set of actions available to be performed by the agent within the environment;
• implement a first machine learning model and a second machine learning model different than the first machine learning model, at least one of the first machine learning model or the second machine learning model being associated with a set of hyperparameters, the set of hyperparameters including at least one of (1) a discount factor associated with a weight of future rewards within the environment or (2) a threshold value to balance between exploration and exploitation within the environment;
• provide an input to the first machine learning model, the input to the first machine learning model based on the first value associated with the property of the environment, the first machine learning model configured to generate an output based on the input to the first machine learning model;
• provide an input to the second machine learning model, the input to the second machine learning model based on the first value associated with the property of the environment, the second machine learning model configured to generate an output based on the input to the second machine learning model;
• select, based on the output of the first machine learning model and the output of the second machine learning model, an identified action from the set of actions available to be performed by the agent within the environment, the identified action being available to be performed when the agent is at the first state and the identified action configured to induce the agent to transition from the first state to a second state from the set of states, the second state predictive of a second value associated with the property of the environment;
• and
• send instructions to the agent to perform the identified action.

A non-transitory processor-readable medium storing code representing instructions to be executed by a processor, the instructions comprising code to cause the processor to:

• deploy an agent in an environment, the agent configured to interact with the environment by performing at least one action selected from a set of actions available to be performed by the agent within the environment;
• receive a value associated with a property of the environment when the agent is at a first state from a set of states defined within the environment;
• provide an input to a machine learning model, the input based on the value associated with the property of the environment, the machine learning model being associated with a set of hyperparameters, the set of hyperparameters including at least one of (1) a discount factor associated with a weight of future rewards within the environment or (2) a threshold value to balance between exploration and exploitation within the environment, the machine learning model configured to generate an output based on the input and the set of hyperparameters;
• select, based on the output, one or more identified actions from the set of actions available to be performed by the agent within the environment, the one or more identified actions being available to be performed when the agent is at the first state and the one or more identified actions configured to induce the agent to transition from the first state to a second state from the set of states;
• and
• send instructions to the agent to perform the one or more identified actions in the environment.

A method, comprising:

• receiving a value associated with a property of an environment in which an agent is deployed at a first state from a set of states defined within the environment, the agent configured to interact with the environment by performing at least one action selected from a set of actions available to be performed by the agent within the environment;
• provide an input to a machine learning model, the input based on the value associated with the property of the environment, the machine learning model being associated with a set of hyperparameters, the set of hyperparameters including at least one of (1) a discount factor associated with a weight of future rewards within the environment or (2) a threshold value to balance between exploration and exploitation within the environment, the machine learning model configured to generate an output based on the input and the set of hyperparameters, the machine learning model trained using a synthetic state generated based on data associated with past actions of the agent in the environment and outcomes associated with the past actions of the agent in the environment;
• selecting, based on the output, one or more identified actions from the set of actions available to be performed by the agent within the environment, the one or more identified actions configured to cause a change in the property of the environment;
• and
• sending instructions to the agent to perform the one or more identified actions in the environment.