Reinforcement Learning for Combining Search Methods in the Calibration of Economic ABMs
Università Cattolica del Sacro Cuore(圣心天主教大学);Universität Bielefeld(比勒费尔德大学)
Aldo Glielmo, Marco Favorito, Debmallya Chanda, Domenico Delli Gatti
备注:
9 pages and 5 figures, presented at the AAAI bridge program 'AI for Financial Institutions' (https://aaai23.bankit.art/), at the ICLR bridge program 'AI4ABM' (https://ai4abm.org/workshop_iclr2023/) and at ICAIF '23 (https://ai-finance.org/). Proceedings of the Fourth ACM International Conference on AI in Finance, (ICAIF 23), Association for Computing Machinery, New York, NY, USA
Calibrating agent-based models (ABMs) in economics and finance typically involves a derivative-free search in a very large parameter space. In this work, we benchmark a number of search methods in the calibration of a well-known macroeconomic ABM on real data, and further assess the performance of "mixed strategies" made by combining different methods. We find that methods based on random-forest surrogates are particularly efficient, and that combining search methods generally increases performance since the biases of any single method are mitigated. Moving from these observations, we propose a reinforcement learning (RL) scheme to automatically select and combine search methods on-the-fly during a calibration run. The RL agent keeps exploiting a specific method only as long as this keeps performing well, but explores new strategies when the specific method reaches a performance plateau. The resulting RL search scheme outperforms any other method or method combination tested, and does not rely on any prior information or trial and error procedure.
Surrogate Assisted Methods for the Parameterisation of Agent-Based Models
University of the Witwatersrand(威特沃特斯兰德大学);University of Johannesburg(约翰内斯堡大学)
参数校准是基于主体的建模与仿真(ABMS)中的一个主要挑战。随着基于代理的模型(ABM)的复杂性的增加,需要校准的参数的数量也在增加。这导致ABMS等同于\say{curse of dimensionality}。我们提出了一个ABMS框架,有利于有效地整合不同的采样方法和代理模型(SM),以评估这些策略如何影响参数校准和探索。我们表明,替代辅助方法比标准的抽样方法更好地执行。此外,我们表明,XGBoost和决策树SM是最优的整体方面,我们的分析。
Calibrating Agent-based Models to Microdata with Graph Neural Networks
Institute for New Economic Thinking(新经济思维研究所);University of Oxford(牛津大学);Improbable(Impossible公司(注:该公司为英国科技公司,通常保留原名Improbable));Santa Fe Institute(圣塔菲研究所);Durham University(杜伦大学)
Generating Origin-Destination Matrices in Neural Spatial Interaction Models
Department of Engineering, Cambridge University(剑桥大学工程系);The Alan Turing Institute(艾伦·图灵研究所);Departments of Statistics & Computer Science, University of Warwick(沃里克大学统计学与计算机科学系)