Calibrating Agent-Based Financial Markets Simulators with Pretrainable Automatic Posterior Transformation-Based Surrogates

TL;DR

This work tackles the high computational cost of calibrating financial agent-based market simulators by introducing ANTR, a surrogate-assisted framework that models the parameter posterior $p(\boldsymbol{\theta}\mid \boldsymbol{x}_{\mathrm{obs}})$ via Automatic Posterior Transformation (APT). By coupling NCS-based diversity with an adaptive trust-region scheme, ANTR maintains exploration while concentrating expensive simulations in promising regions, achieving better parameter recovery and efficiency than state-of-the-art SAEAs. Empirical results on the Brock–Hommes and PGPS benchmarks demonstrate superior parameter-level calibration, strong batch-transfer ability, and robustness to varying sequence lengths, suggesting practical value for policy analysis and risk assessment. The approach advances simulation-based inference for ABMs by leveraging pretrained posterior surrogates to bridge data-space discrepancies and parameter-space accuracy, enabling scalable, batch-calibrated market simulations.

Abstract

Calibrating Agent-Based Models (ABMs) is an important optimization problem for simulating the complex social systems, where the goal is to identify the optimal parameter of a given ABM by minimizing the discrepancy between the simulated data and the real-world observations. Unfortunately, it suffers from the extensive computational costs of iterative evaluations, which involves the expensive simulation with the candidate parameter. While Surrogate-Assisted Evolutionary Algorithms (SAEAs) have been widely adopted to alleviate the computational burden, existing methods face two key limitations: 1) surrogating the original evaluation function is hard due the nonlinear yet multi-modal nature of the ABMs, and 2) the commonly used surrogates cannot share the optimization experience among multiple calibration tasks, making the batched calibration less effective. To address these issues, this work proposes Automatic posterior transformation with Negatively Correlated Search and Adaptive Trust-Region (ANTR). ANTR first replaces the traditional surrogates with a pretrainable neural density estimator that directly models the posterior distribution of the parameters given observed data, thereby aligning the optimization objective with parameter-space accuracy. Furthermore, we incorporate a diversity-preserving search strategy to prevent premature convergence and an adaptive trust-region method to efficiently allocate computational resources. We take two representative ABM-based financial market simulators as the test bench as due to the high non-linearity. Experiments demonstrate that the proposed ANTR significantly outperforms conventional metaheuristics and state-of-the-art SAEAs in both calibration accuracy and computational efficiency, particularly in batch calibration scenarios across multiple market conditions.

Figures (4)

• Figure 1: Overall framework of the proposed ANTR framework
• Figure 2: Illustration of the Negatively Correlated Search (NCS) framework. The algorithm maintains $N$ parallel RLS processes, each of which can be regarded as a probability distribution $p_i(\boldsymbol{\theta})$.
• Figure 3: Comparison of MSE convergence curves on the Brock and Hommes Heterogeneous Expectations Model for ANTR, NCS, TuRBO, and CAL-SAPSO.
• Figure 4: Comparison of MSE convergence curves on the MAXE model for ANTR, NCS, TuRBO, and CAL-SAPSO.