Cost-Optimal Foundation Model Deployment Portfolio for Transportation Management
Quick Answer
The study presents the Foundation Model Deployment Portfolio (FMDP) problem for optimizing model deployment in transportation management centers, achieving a cost of $34/month—97% lower than the all-closed-API baseline—by utilizing open-source APIs for four functions.
Quick Take
The study presents the Foundation Model Deployment Portfolio (FMDP) problem for optimizing model deployment in transportation management centers, achieving a cost of $34/month—97% lower than the all-closed-API baseline—by utilizing open-source APIs for four functions. A polynomial-time greedy heuristic is proposed, with break-even analysis indicating on-premise GPU investment is viable only above 309 vision queries/hour or if API prices double.
Key Points
- FMDP minimizes total cost of ownership while meeting quality and latency constraints.
- The problem is proven NP-hard, reducing from the 0-1 knapsack problem.
- A case study with five TMC functions identified a mixed portfolio costing $34/month.
- Four functions were routed to open-source APIs, one to a closed API.
- On-premise GPU investment is reasonable only above 309 vision queries/hour.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Foundation models, including large language models (LLMs) and vision-language models (VLMs), are increasingly used for transportation management center (TMC) tasks such as anomaly detection, incident reporting, and traveler information. Deploying multiple such models across TMC functions raises a portfolio question: which model should serve each function, in which deployment mode, and under what shared hardware budget? We formulate this as the Foundation Model Deployment Portfolio (FMDP) problem, a mixed-integer program minimizing total cost of ownership (TCO) subject to per-function quality, latency, and safety constraints over shared GPU capacity. We prove the problem NP-hard by reduction from the 0-1 knapsack problem and propose a polynomial-time greedy heuristic. In an illustrative case study with five TMC functions and 19 candidate (model, mode) pairs, FMDP identifies a mixed portfolio costing $34/mo (97% below the cheapest feasible all-closed-API baseline) by routing four functions to open-source APIs and the one function whose quality floor no open-source model meets to a closed API. Break-even analysis shows that on-premise GPU investment becomes reasonable only above approximately 309 vision queries/hour or if API prices double.
| Comments: | Accepted at IEEE ITSC 2026 |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2607.13239 [cs.AI] |
| (or arXiv:2607.13239v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2607.13239 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Xi Cheng [view email]
[v1]
Tue, 14 Jul 2026 20:01:15 UTC (18 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →Automatic Ordinary Differential Equations Discovery For Biological Systems Using Large Language Model Powered Agentic System
The MEDA system utilizes large language models and symbolic regression to autonomously discover ordinary differential equations for biological systems, achieving strong structural recovery and biologically plausible models. It outperforms existing methods by integrating domain knowledge and mechanistic constraints, demonstrating effective retrieval and extrapolation capabilities.
