System and methods for creation of learning agents in simulated environments
A system and methods for generating and applying learning agents in simulated environments, in which an agent simulation is selected, one or more agent goals are received, and agents are created which are individual instances of the agent simulation with each agent having at least one of the agent goals, wherein the agents are used in the execution of an environment simulation which dynamically changes based on the collective behavior of the agents.
1. A system for generating learning agents in simulated environments, comprising:
a computing device comprising a memory and a processor;
an agent creation engine comprising a first plurality of programming instructions stored in the memory and operating on the processor, wherein the first plurality of programming instructions, when operating on the processor, cause the computing device to:
receive one or more agent goals;
select an agent simulation based on the one or more agent goals;
create a plurality of agents, each agent being an instance of the agent simulation and having one of the one or more goals; and
at least one simulation manager comprising a second plurality of programming instructions stored in the memory and operating on the processor, wherein the second plurality of programming instructions, when operating on the processor, cause the computing device to:
receive a simulation goal related to the one or more agent goals;
select a dynamic environment simulation based on the simulation goal;
execute the dynamic environment simulation using a plurality of meta-models, wherein the plurality of simulation agents used in the execution is managed according to at least one meta-model, wherein the at least one meta-model comprises a plurality of relationships between the dynamic environment simulation and the agent simulation; and
continue execution of the dynamic environment simulation that evolves with agent behavior from the execution of the plurality of agents and the at least one meta-model until the simulation goal has been reached or until each agent has achieved at least one agent goal from the one or more agent goals;
wherein each of the plurality of agents takes different actions in a non-deterministic environment based on specifications of the respective agent and specifications of the dynamic environment simulation to achieve the one or more agent goals; wherein the actions taken by the agents include hierarchical learning according to set agent goals, learning patterns of user behavior and learned behavior of agents stored from previous simulations;
wherein the different actions and learned behaviors acquired by individual agents further differentiating them from each other during the simulation execution; and
examine the outcomes of each individual set of actions taken by each agent according to the simulation goal.
2. The system of claim 1 , wherein the agent creation engine operates on a simulation execution server.
3. The system of claim 1 , wherein the agent creation engine operates on a separate server from the simulation execution server.