Population Synthesis

Population synthesis aims to create realistic, synthetic populations of agents for simulations, mirroring real-world demographic and behavioral characteristics. Current research focuses on improving the accuracy and scalability of these synthetic populations, employing techniques like Bayesian networks, variational autoencoders, generative adversarial networks, and multi-objective optimization to capture complex relationships between individual and household attributes, including newly incorporated motivational factors. These advancements enhance the realism and utility of agent-based models across diverse fields, from urban planning and policy analysis to astrophysics and medical device testing, by providing more accurate and representative input data.