Package com.tnsai.agents.ensemble

Class AgentEnsembleExecutor

java.lang.Object

com.tnsai.agents.ensemble.AgentEnsembleExecutor

public class AgentEnsembleExecutor
extends Object

Handles multi-LLM ensemble operations for an Agent.

This class extracts ensemble/parallel LLM concerns from the Agent class:

• Parallel fan-out to multiple LLMs
• Race pattern (first response wins)
• Consensus/scoring based selection
• Majority voting
• Ensemble synthesis with aggregator LLM

Constructor Summary

Constructors

Constructor	Description
AgentEnsembleExecutor()

Method Summary

Modifier and Type	Method	Description
Optional<ParallelLLM.LLMResult>	consensusChat(String message, List<LLMClient> llms, Function<String,Double> scorer)	Sends message to multiple LLMs and selects the best response using a scoring function.
Optional<String>	ensembleChat(String message, List<LLMClient> llms, LLMClient aggregator)	Sends message to multiple LLMs and combines responses using an aggregator LLM.
Optional<ParallelLLM.LLMResult>	majorityVoteChat(String message, List<LLMClient> llms)	Sends message to multiple LLMs and selects by majority voting.
ParallelLLM.ParallelResults	parallelChat(String message, List<LLMClient> llms)	Sends message to multiple LLMs in parallel and collects all responses.
Optional<ParallelLLM.LLMResult>	raceChat(String message, List<LLMClient> llms, long timeoutSeconds)	Sends message to multiple LLMs and returns the first successful response.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

AgentEnsembleExecutor

public AgentEnsembleExecutor()

Method Details

parallelChat

public ParallelLLM.ParallelResults parallelChat(String message,
 List<LLMClient> llms)

Sends message to multiple LLMs in parallel and collects all responses.

Parameters:

message - The message to send

llms - List of LLM clients to query in parallel

Returns:

Aggregated results from all LLMs

raceChat

public Optional<ParallelLLM.LLMResult> raceChat(String message,
 List<LLMClient> llms,
 long timeoutSeconds)

Sends message to multiple LLMs and returns the first successful response.

Parameters:

message - The message to send

llms - List of LLM clients to race

timeoutSeconds - Timeout in seconds

Returns:

First successful response, or empty if all fail

consensusChat

public Optional<ParallelLLM.LLMResult> consensusChat(String message,
 List<LLMClient> llms,
 Function<String,Double> scorer)

Sends message to multiple LLMs and selects the best response using a scoring function.

Parameters:

message - The message to send

llms - List of LLM clients

scorer - Function to score each response (higher is better)

Returns:

Best response based on scoring

majorityVoteChat

public Optional<ParallelLLM.LLMResult> majorityVoteChat(String message,
 List<LLMClient> llms)

Sends message to multiple LLMs and selects by majority voting.

Parameters:

message - The message to send

llms - List of LLM clients

Returns:

Most common response pattern

ensembleChat

public Optional<String> ensembleChat(String message,
 List<LLMClient> llms,
 LLMClient aggregator)

Sends message to multiple LLMs and combines responses using an aggregator LLM.

Parameters:

message - The message to send

llms - List of LLM clients for initial responses

aggregator - LLM client to synthesize final response

Returns:

Synthesized response combining best elements