This study examines the potential of large language models (LLMs) to substitute for human respondents in survey research. Surveys serve as essential tools in fields like social science, marketing, and policy-making; however, traditional methods often require considerable time and costs. LLMs present a promising alternative to mitigate these burdens, though their reliability—particularly outside of U.S. contexts—remains uncertain. This study focuses on surveys conducted in Japan, comparing the responses generated by LLMs to those of actual Japanese participants. Our analysis reveals notable discrepancies due to inherent biases in LLMs, though adjusting the models to better align with specific personas can partially enhance the accuracy of simulated responses. We emphasize the need for further research to fully understand the capabilities and limitations of LLMs, aiming to refine their application in diverse areas such as social sciences, marketing, and policy decision-making.

This study examines the timing and extent of congestion in situations where agents with a continuous distribution of target arrival times make decisions about their departure times. In most real-world scenarios, arrival times are heterogeneous, and research that sufficiently considers both heterogeneity and uncertainty has been lacking. In this study, we formulated the real-world traffic situation as an optimisation problem that accounts for both the heterogeneity and uncertainty of arrival times in various scenarios. The results indicate that effective monetary incentives for avoiding congestion require careful pricing. Appropriately designed incentives are expected to alleviate road congestion. Policymakers can conduct a preliminarily evaluation of their impact using the approach employed in this study.

This study explores the optimal settings for a prevalent simulation in which agents powered by large language models (LLMs) make decisions without predetermined actions as a substitute for economic experiments. Economic experiments are essential methods in economics where the behaviors of participants are observed under controlled conditions to test hypotheses and theories, involving significant time, effort, and cost. If a simulation can supplant economic experiments, researchers can overcome these limitations by utilizing it. We focused on three essential factors for applying the simulation using LLMs to economic experiments and conducted simulations for both the proposer and responder sides of the one-shot ultimatum game in various settings of the three factors. The sensitivity analysis revealed that, on the proposer side, there was a setting that produced results similar to actual human-centered experiments. However, on the responder side, none of the simulation setups yielded results consistent with real-world experimental data. These findings suggest both the potential and limitations of the novel simulation with LLMs as a substitute for economic experiments.

In this study, we evaluated the impact of introducing a new transportation system, bus rapid transit (BRT), on traffic congestion using a multi-agent simulation (MAS). The rapid urbanisation accompanying the development of cities worldwide, with a growing population of over one million people living in cities, is expected to result in increased traffic congestion. Therefore, addressing this problem has become increasingly urgent. One of the potential solutions is the BRT system, which has attracted considerable interest. However, research on its effectiveness in alleviating congestion has not been sufficiently conducted, which was the focus of this study.We conducted MAS using the tool Simulation of Urban MObility (SUMO), focusing on the Eastern Saitama region as a case study and evaluated the impact of congestion relief resulting from the introduction of BRT.We observed that the introduction of BRT had a positive impact on increasing the maximum traffic flow rate and contributed to congestion relief through the adjustment of BRT headways. Additionally, we obtained suggestions for further congestion relief by designing of dedicated BRT lanes. The outcomes of this study can be leveraged to provide strategic recommendations and decision-making support for road construction and BRT introduction plans.

To extend people’s life expectancy while improving their health and quality of life, the promotion of artificial intelligence (AI) and big data technologies to life science fields is crucial. In the medical field, AI model training is quite difficult owing to its heterogeneity and the fact that the stakeholders and data owners have a variety of valid privacy and protection needs and concerns, which has resulted in heavily suppressed cooperation. To help remedy this problem, we examine the behaviors of life science researchers who use medical data for AI purposes from social and institutional perspectives. From this, we provide a mathematical model for medical data provisioning based on the public goods game theory. Numerical simulation of the proposed model shows that the "pump-priming effect" caused by the existence of voluntary data providers may result in cooperative behaviors instead of the non-cooperative equilibrium obtained in the conventional public goods game. We conclude that our approach provides an effective mathematical and quantitative mechanism design in response to extant qualitative medical data provisioning problems.