Large Language Models (LLMs) show remarkable potential for urban computing, from spatial reasoning to predictive analytics. However, evaluating LLMs across diverse urban tasks faces two critical challenges: lack of unified platforms for consistent multi-source data access, and fragmented task definitions that hinder fair comparison. To address these challenges, we present CityVerse, the first unified platform integrating multi-source urban data, capability-based task taxonomy, and dynamic simulation for systematic LLM evaluation in urban contexts. CityVerse provides: (i) coordinate-based Data APIs unifying ten categories of urban data—including spatial features, temporal dynamics, demographics, and multi-modal imagery-with over 38 million curated records; (ii) Task APIs organizing 43 urban computing tasks into a four-level cognitive hierarchy: Perception, Spatial Understanding, Reasoning Prediction, and Decision Interaction, enabling standardized evaluation across capability levels; (iii) an interactive visualization frontend supporting real-time data retrieval, multi-layer display, and simulation replay for intuitive exploration and validation. We validate the platform's effectiveness through evaluations on mainstream LLMs across representative tasks, demonstrating its capability to support reproducible and systematic assessment. CityVerse provides a reusable foundation for advancing LLMs and multi-task approaches in the urban computing domain. The framework is publicly available at https://julietjobs.github.io/City_Verse/.

The electrification of shared mobility has become popular across the globe. Many cities have their new shared e-mobility systems deployed, with continuously expanding coverage from central areas to the city edges. A key challenge in the operation of these systems is fleet rebalancing, i.e., how EVs should be repositioned to better satisfy future demand. This is particularly challenging in the context of expanding systems, because i) the range of the EVs is limited while charging time is typically long, which constrain the viable rebalancing operations; and ii) the EV stations in the system are dynamically changing, i.e., the legitimate targets for rebalancing operations can vary over time. We tackle these challenges by first investigating rich sets of data collected from a real-world shared e-mobility system for one year, analyzing the operation model, usage patterns and expansion dynamics of this new mobility mode. With the learned knowledge we design a high-fidelity simulator, which is able to abstract key operation details of EV sharing at fine granularity. Then we model the rebalancing task for shared e-mobility systems under continuous expansion as a Multi-Agent Reinforcement Learning (MARL) problem, which directly takes the range and charging properties of the EVs into account. We further propose a novel policy optimization approach with action cascading, which is able to cope with the expansion dynamics and solve the formulated MARL. We evaluate the proposed approach extensively, and experimental results show that our approach outperforms the state-of-the-art, offering significant performance gain in both satisfied demand and net revenue.

For many scientific disciplines, the continued progression of information technology has increased the availability of data, computation and analytical methodologies including simulation and visualisation. Geographical information science is no exception. In this article, we investigate the possibilities for deployment of e-infrastructures to inform spatial planning, analysis and policy-making. We describe an existing architecture that feeds both static and dynamic simulation models from a variety of sources, including not only administrative datasets but also attitudes and behaviours which are harvested online from crowds. This infrastructure also supports visualisation and computationally intensive processing. The main aim of this article is to illustrate how spatial simulation models can be calibrated with crowd-sourced data. We introduce an example in which popular attitudes to congestion charging in a major UK city (Manchester) were collected, with promotional support from a high-profile media organisation (the BBC). These data are used to estimate the parameters of a transport simulation model, using a hungry estimation procedure which is deployed within a high-performance computational grid. We indicate how the resulting model might be used to evaluate the impact of alternative policy options for regulating the traffic in Manchester. Whilst the procedure is novel in itself, we argue that greater credibility could be added by the incorporation of open-source simulation models and by the use of social networking mechanisms to share policy evaluations much more widely.

In a collaborative project between GMAP Ltd and EPCC, an existing heuristic optimisation scheme for strategic resource planning was parallelised to run on the data parallel Connection Machine CM-200. The parallel software was found to run over 2700 times faster than the original workstation software. This has allowed the exploration of complex business planning strategies at a national, rather than regional, level for the first time. The availability of a very fast evaluation program for planning solutions also enabled an investigation of the use of genetic algorithms in place of GMAP's existing heuristic optimisation scheme. The results of this study show that genetic algorithms can provide better quality solutions in terms of both predicted profit from the solution and spatial diversity to provide a range of possible solutions. This paper discusses both the parallelisation of the original optimisation scheme and the use of genetic algorithms in place of this method. © 1997 John Wiley & Sons, Ltd.