Aligning LLM agents with human learning and adjustment behavior: a dual agent approach

TL;DR

This work tackles the challenge of accurately simulating how human travelers learn and adjust their route choices over time. It introduces a dual-LLM framework with memory-augmented traveler agents and an online LLM calibration agent that uses a textual pseudo-gradient descent to adapt agent personas from online data. The approach yields superior alignment at both the individual level and the aggregate system dynamics compared with several baselines, and it demonstrates the model's ability to capture evolving learning processes for robust generalization. The framework offers a scalable, API-friendly pathway to behaviorally realistic transportation simulations, with implications for policy analysis and planning under dynamic conditions.

Abstract

Effective modeling of how human travelers learn and adjust their travel behavior from interacting with transportation systems is critical for system assessment and planning. However, this task is also difficult due to the complex cognition and decision-making involved in such behavior. Recent research has begun to leverage Large Language Model (LLM) agents for this task. Building on this, we introduce a novel dual-agent framework that enables continuous learning and alignment between LLM agents and human travelers on learning and adaptation behavior from online data streams. Our approach involves a set of LLM traveler agents, equipped with a memory system and a learnable persona, which serve as simulators for human travelers. To ensure behavioral alignment, we introduce an LLM calibration agent that leverages the reasoning and analytical capabilities of LLMs to train the personas of these traveler agents. Working together, this dual-agent system is designed to track and align the underlying decision-making mechanisms of travelers and produce realistic, adaptive simulations. Using a real-world dataset from a day-to-day route choice experiment, we show our approach significantly outperforms existing LLM-based methods in both individual behavioral alignment and aggregate simulation accuracy. Furthermore, we demonstrate that our method moves beyond simple behavioral mimicry to capture the evolution of underlying learning processes, a deeper alignment that fosters robust generalization. Overall, our framework provides a new approach for creating adaptive and behaviorally realistic agents to simulate travelers' learning and adaptation that can benefit transportation simulation and policy analysis.

Figures (6)

• Figure 1: Overview of our approach
• Figure 2: Components and workflow of an LLM traveler agent
• Figure 3: Components and workflow of the LLM calibration agent
• Figure 4: Global behavior vector distribution (Day 20–39) across methods. Marker color intensity reflects cosine similarity to the corresponding human traveler.
• Figure 5: Simulated route flows under controlled system dynamics