AgenticMobility: An Open-Source Platform for Multi-AI-Agent Mobility Dynamics
Authors: Jiaheng Xiao, Jiangtao Han, Jiachen Wang, Honghai Li, Lijun Sun, Kay Axhausen, Chang Liu, Zhengbing He
Organizations: Faculty of Science and Engineering, University of Nottingham Ningbo China, China · College of Management and Economics, Tianjin University, China · Beijing Key Laboratory of Traffic Engineering, Beijing University of Technology, China · Research Institute of Highway, Ministry of Transport, China · Department of Civil Engineering, McGill University, Canada · Institute for Transport Planning and Systems, ETH Zurich, Switzerland · School of Economics, University of Nottingham Ningbo China, China
Understanding mobility dynamics and the effects of urban transportation management policies requires examining how individual travel decisions, traffic conditions, and policy interventions influence one another over time. We present AgenticMobility, an open-source experimental platform for studying how mobility AI agents and other AI agents interact, adapt through repeated experience, and collectively shape mobility outcomes. Persistent agents retain identities and relevant states across rounds, with explicit interfaces defining information access, admissible actions, and decision schedules; linked records trace decisions, execution, and feedback. In the current implementation, large language model (LLM) agents act as travelers choosing routes and departure times and as a manager periodically adjusting link charges within MATSim. A preliminary demonstration spans 100 simulated days in each of five configurations, with one run per configuration and a shared simulation seed. In the joint traveler--manager run, all 1,179,300 finalized target decisions satisfied action constraints; execution checks verified implementation of accepted choices and preservation of personal memories across days. Trajectories show stable route choices alongside departure-time adjustments and differences between controlled-trip and population-wide outcomes. The platform supports experiments on individual adaptation, collective outcomes, and management interventions by varying information access, retained experience, and management feedback. Source code is available at https://github.com/Agentic-City-Research/Agentic-Mobility
Figures & tables
Figure 1: AgenticMobility interaction framework. Role-specific actions and feedback connect agents through a shared execution environment. Persistent states link successive rounds; a separate archive records the interaction process. The traveler and manager roles illustrate one study setting rather than a fixed set of platform roles.
Interface
Required information
Observation
Receiving agent and role, decision round, information sources, and access permissions.
Action
Acting agent, submitted values, affected entities or resources, effective timing, and validity constraints.
Outcome
Relevant actions and execution round, application and completion status, and realized results.
State update
Previous persistent state, newly available feedback, and updated state.
Table 1: Information requirements for the platform interfaces. Entries specify the meaning and scope of exchanged information; implementation-specific fields may vary.
Figure 2: Illustrative interaction timing and state updates across successive rounds for traveler and manager roles. Solid arrows indicate action and feedback flows; dashed arrows indicate state transfer and research-record linkage. Role schedules are study-specific, and the research archive remains separate from agent-accessible memory.
Traveler
Manager
Component
Mechanism
Setting
Mechanism
Setting
Role
Target-trip decision maker
11,793 agents
Road pricing manager
1 agent
Action
Route choice; departure-time adjustment
≤4 routes; −90 to +30 min, 15-min steps
Link-charge adjustment
≤40 car links
Observation
Personal trip outcomes and permitted toll information
—
Public network conditions and aggregate outcomes
—
Memory
Factual history; personal-experience summary
3-day history; 500-character summary
Policy history; public-feedback summary
Summary ≤1,000 words
Decision schedule
Trip decisions
Daily
Policy reviews
Days 20, 50, and 80
Table 2: Role-specific agent mechanisms and settings in AgenticMobility.
Task
Model identifier
Output format
Max. tokens
Traveler choice
qwen3.5-9b
JSON schema
1,024
Personal memory
qwen3.5-9b
JSON schema
1,024
Manager policy
deepseek-flash
JSON object
8,192
Manager memory
deepseek-flash
JSON object
2,048
Table 3: Recorded model settings for LLM Traveler + Manager. Token limits are per-response maxima.
Figure 3: Experience retention and state updates. New outcomes update the rolling factual history and supply summary generation. Validated summaries enter later inputs. Manager memory updates daily, independently of policy decisions, and the research archive remains outside unrestricted agent access.
Figure 4: Components and data flows in the implementation. The coordinator manages agent decisions, validation, and state updates, while the execution adapter connects these mechanisms to the MATSim backend. Returned events support role-specific feedback, with agent states and research records maintained separately.
Material
Intended use
Source code, scenario inputs, and configuration guides
Build and run the platform and configure new experiments.
Aggregate outputs and run summaries
Inspect daily trajectories and recalculate reported summary measures.
Execution-audit records and selected case excerpts
Inspect recorded action checks and examples linking decision inputs, actions, and outcomes.
Run metadata and artifact hashes
Identify the recorded runs and trace available executable and input artifacts.
Table 4: Materials provided with the AgenticMobility release and their intended uses.
Configuration
Target-trip decisions
LLM manager
MATSim
MATSim plan selection and replanning
Disabled
MNL
Constrained joint route–departure choice
Disabled
LLM Traveler
LLM route–departure choice
Disabled
MATSim + Manager
MATSim plan selection and replanning
Enabled
LLM Traveler + Manager
LLM route–departure choice
Enabled
Table 5: Demonstration configurations. Trips outside the controlled target set retain MATSim behavioral mechanisms.
Table 6: Principal settings for the demonstration. Strategy entries are configured weights, not probabilities inferred from their sum.
Check
Coverage
Recorded result
Action constraints and execution
1,179,300 finalized target decisions in the combined run
Valid route and departure choices; accepted actions linked to applied plans and completed routes
Personal state continuity
100-day personal-memory chains in the combined run
Continuous chains; finalized summaries within the configured limit
Manager decisions and state
3 policy decisions and 100 daily memory updates in the combined run
Valid policies; all memory updates completed
Information access
4,000 contexts: 2,000 from each LLM traveler configuration
No current announcements disclosed to uninformed travelers and no network-wide traffic information disclosed
Table 7: Execution audit scope and results. Action checks concern finalized outputs after validation and repair. The information-access result applies to the inspected samples.
Configuration
API calls
Tokens
MATSim
0
0
MNL
0
0
LLM Traveler
2,362,964
5,713,538,990
MATSim + Manager
109
872,774
LLM Traveler + Manager
2,364,090
5,451,720,152
Table 8: LLM inference workload over complete 100-day runs. Calls include retries and repairs; tokens are service-reported totals.
Figure 6: Individual and population mobility outcomes across five configurations: (a) mean daily car time per person, (b) mean target-trip car time, and (c) work-target late-arrival rate. Thin lines show daily values; thick lines show trailing seven-day means. Positive charging begins on Day 20 for MATSim + Manager and Day 50 for LLM + Manager. Day 90 marks the MATSim innovation cutoff. Each configuration has one seed-268 run.
Figure 7: Individual choice changes and aggregate route-flow turnover. (a) Route-switching rates across five configurations on a logarithmic scale. (b) Departure-time changes in the two LLM traveler configurations. (c) Individual route switching and net turnover of OD-specific route counts. Rates use all 11,793 targets. Thin lines show daily values; thick lines show trailing seven-day means. Day 50 marks charging in LLM + Manager, and Day 90 marks the MATSim innovation cutoff.
Appendix figures & tables4 assets
Supplementary material from the paper’s appendix.
Appendix
Stage
Recorded connection
Before Day 50 execution
Public observations through Day 49; empty charging policy; summary written after Day 49
Day 50 policy decision
Accepted test action: 1.5 units per entry on 43_3 and 43_4 over [0,108000) seconds
Day 50 policy application
Active policy identifier pricing-v2-it50
After Day 50 execution
Public feedback enters a complete replacement summary; the memory update leaves the policy unchanged
Day 51 memory input
Contains the summary written after Day 50
Day 80 policy decision
Accepted hold action; policy identifier changes to pricing-v3-it80 , but the price fingerprint is unchanged
Appendix
Table 9: Linked manager records in the combined run. Days use zero-based log indices. A memory update does not itself change the active policy.
Figure 8: Validation and task-specific failure handling. Persistent traveler task failure stops the run. Manager policy output failure retains the active policy, and manager-memory failure retains the previous summary. Runtime and transport failures are handled separately, as described in the text.
Configuration
Daily car time (min/person)
Target time (min/trip)
Work lateness (%)
MATSim
44.34
29.32
19.97
MNL
61.35
53.19
31.25
LLM Traveler
35.52
17.40
5.61
MATSim + Manager
45.16
29.64
20.42
LLM Traveler + Manager
40.71
16.25
3.73
Appendix
Table 10: Mean outcomes over the 100-day window. Daily car time is normalized by 16,822 people; target time covers 11,793 trips; work lateness covers 11,332 work targets.
Configuration
Overall route changes
Multi-route changes
Departure changes
MATSim
5.26
—
20.05
MNL
48.55
63.52
85.24
LLM Traveler
3.08
4.03
29.45
MATSim + Manager
5.43
—
20.28
LLM Traveler + Manager
2.19
2.86
28.04
Appendix
Table 11: Mean choice-change rates over 99 transitions. Overall rates use 11,793 targets; multi-route rates use 9,013 targets with multiple fixed candidates. All values are percentages. A dash denotes a candidate-space comparison that is not applicable.
Agent-based modeling simulates how interacting, heterogeneous entities and their environments produce patterns in social, economic, biological, and engineered systems.
Recent LLM-based multi-agent urban simulators can generate semantically rich city routines, but they remain costly to scale and are often weakly validated against empirical mobility patterns. We present CityBehavEx, an interactive LLM-assisted urban simulation platform that scales to city-size populations, exposes agent behavior for inspection, supports empirical validation, and generates mobility patterns that better match real-world spatial, temporal, and semantic distributions. Instead of invoking large language models for every agent action, CityBehavEx combines established human mobility models with fine-tuned cross-encoders that estimate semantic alignment between agent profiles, schedules, and activity transitions. This design enables large-scale simulations, as demonstrated in a case study of 100,000 agents over 75 days in under one hour on a single consumer GPU. The platform allows users to define simulation regions, launch experiments, inspect trajectories and activity traces, debug unrealistic behaviors, and validate generated routines against real-world mobility, time-use, and semantic metrics.
Gustavo H. Santos, Aline Viana, Thiago H Silva
UTFPR, Brazil · Inria, France · University of Toronto, Canada
Urban mobility modeling faces challenges in representing decision-making in dynamic environments. Although Multi-Agent Systems are widely used, rule-based approaches rely on fixed heuristics that limit adaptive behavior. This work investigates the integration of Large Language Models (LLMs) as decision-making components in multi-agent simulations. We propose a hybrid architecture that connects the GAMA platform to an external LLM-based module through an API, enabling agents to determine whether route replanning is necessary. Rather than replacing routing algorithms, the LLM serves as a decision layer that guides replanning behavior. The approach incorporates persistent memory, allowing past interactions to influence future decisions and promote behavioral consistency. We compare rule-based and LLM-assisted approaches across multiple road-blockage scenarios and population scales. Results indicate that LLM-enabled agents exhibit greater adaptability and contextual awareness, particularly in scenarios with higher route flexibility. Memory influences performance and behavioral consistency, with effects varying across configurations. Overall, LLMs serve as complementary cognitive layers that enrich behavioral representations in urban mobility simulations and hold potential for modeling complex decision-making in spatial multi-agent systems.
Postgraduate Program in Computing Federal University of Pelotas, Pelotas, Brazil · Undergraduate in Computer Science Federal University of Pelotas, Pelotas, Brazil · Undergraduate in Computer Engineering Federal University of Pelotas, Pelotas, Brazil
In existing agent-based simulations of multimodal urban mobility, agents, representing humans, choose their travel modes either using discrete choice or machine learning models fitted on structured travel survey data, or with rules explicitly written by the domain experts. Such approaches fix a priori the agents' possible behaviour space. Hence, a factor absent from variables or rules cannot affect agents' decisions. In reality, choosing a transport mode is a multidimensional decision in which objective and subjective factors intertwine and change over time. Integrating such complexity into a city-scale multimodal mobility simulation remains a challenge. Generative agents (GAs) built on large language models promise to respond to this challenge because they carry decision heuristics that travel surveys do not record. However, to the best of our knowledge, aligning their decisions with real household travel surveys remains an open question. In this paper, we build an agent-based model of daily multimodal mobility for Toulouse (France). We compare the performance of fifteen decision-makers (simple baselines, models fitted on the survey, GAs with minimal and tuned prompts, and a typed classifier) on the real household travel survey. Our results show that GAs with tuned prompts come close to the fitted models on the modal split of a simulated day without transport disruptions, but not trip by trip. GAs also react when unforeseen events occur. The recently released Jev-1.13 TypeSafe classifier gives surprisingly great results, showing that generating output text is not necessary. The results presented in this paper would be of interest to transport policy makers and service providers looking to apply generative agent-based modelling for improving urban mobility services.
Yves Bru, Kamaldeep Singh Oberoi, Benoit Gaudou
UMR IRIT, Université Toulouse Capitole, Toulouse, France · CESI, CESI LINEACT, Toulouse, France