What we're building
The system has two parts. First, a set of small tools, each trained to do one job in the network, like saving energy or raising throughput. Second, an AI on top, the orchestrator, that learns which tools to use and in what order, based on what the network needs at that moment.
01 · Background
One of us built a photo-restoration project last year. It used a set of small neural networks: one to remove noise, one to remove blur, one to sharpen. An AI decided the order to run them in, and that order changed depending on how the photo was damaged. This project takes the same setup and applies it to a 5G network.
A damaged photo, fixed by a sequence of small tools:
A congested or wasteful network, fixed the same way:
02 · The system
The orchestrator looks at the current state of the network, picks one tool, applies it, and checks what happened. Then it repeats. That loop is the core of the design, and everything below sits inside it.
03 · The tools
Each tool does one job and is trained on its own. We start with the first two, get the idea working, then add the others. Adding a tool does not mean retraining the ones already there, and showing that we can do this is one of the results we want.
Turns a base station off when it isn't needed. Saves a lot of energy, but it can leave a coverage gap.
Raises or lowers a cell's signal strength. More power means better coverage and throughput, and higher energy use.
Moves users from one cell to a neighbour, either to balance load or to clear a cell before it sleeps.
Assigns chunks of radio spectrum to users. This is the main control for throughput and fairness.
Do nothing this step. It matters more than it looks: letting the AI pick “wait” stops it from changing things constantly.
The hard part
The tools work in different ways. One is a switch, one is a dial, one is an assignment. Getting a single AI to control all of them together is the part that existing work handles poorly.
04 · The orchestrator
The orchestrator is a reinforcement-learning agent. We give it the goal as a set of weights, say 70% energy and 30% throughput, and it learns which tool to use for a given network state. Since it works through named tools, each decision reads as a short sequence that an engineer can follow.
Example: a quiet cell at 2 a.m., goal set to save energy
A single black-box model would just output something like “apply config #4471.” This tells you what it changed, and in what order.
05 · What we claim
The claim is not about beating everyone on energy savings. Plenty of papers already do that. It is about two things that rarely show up together.
CLAIM 01
Give the frozen tools a new goal, like prioritising throughput today, or a new network layout, like a denser set of cells, and the orchestrator adjusts without retraining any of the tools. We measure how much of the performance this keeps compared with rebuilding from scratch.
CLAIM 02
Every action is a named tool in a sequence you can see. An operator can check it and debug it instead of trusting a black box. We will show this directly, for example a fault that the decision trace points straight to.
06 · For the telecom folks
Yes. It lines up with O-RAN, the open 5G architecture operators are building now. Our tools are xApps. Our orchestrator sits in the coordination layer, and handling conflicts between xApps is a problem O-RAN itself lists as open. We build the simulator's interfaces to match the real ones, E2 and A1, so moving from the simulator to a real controller means replacing one adapter rather than rewriting the system.
ns-O-RAN, a standards-based simulator.07 · What's new
We read through the 2025 and 2026 papers carefully, partly because a close idea (co-trained hierarchies for O-RAN) came out in December 2025. Here is where our work differs from what already exists.
In one sentence: an orchestrator that handles conflicts between a fixed set of different tools, recombines them for new goals and network layouts without retraining, and explains what it did.
08 · Roadmap
The scope is large, so the order matters. By M2 we already have a complete, publishable result. The harder parts, like the graph neural network for new layouts, come after that, once the safe result is done.
ns-O-RAN.09 · The team (8 of us)
Everyone owns one specific piece, so nothing falls through the cracks. Here is who does what.
Owns the radio & KPI model
Owns the O-RAN mapping & traffic realism
Owns the base orchestrator (RL agent)
Owns the simulator core
Owns the NetworkBackend seam & ns-O-RAN port
ns-O-RAN version that proves sim→real is just an adapter swap.Owns the tools
Owns robustness & security
Owns the graph neural network & the generalization study
10 · Risks
These are the things most likely to cause trouble, and how the plan handles them.
ns-O-RAN version is how we make that gap smaller.