RANOpt AI — Interactive AI-Assisted 5G/6G RAN Decision Platform Beta v17.7

SimulationNot simulated

ConfigurationAwaiting first run

Display modeNormal

AIPPO Pretrained

Deployment Network geometry

Duplex configuration

Inter-site distance (m)

Active UEs / cell

Traffic profile

Penetration loss (dB)

Interference margin (dB)

UE capability Constrains all layers

UE profile

Max DL layers

Max UL layers

UE receiver noise figure (dB)

Coverage FDD layer Anchor layer

Deployment morphology / scenario

5G NR FDD band

DL center frequency (MHz)

Paired UL center frequency (MHz)

DL bandwidth (MHz)

UL bandwidth (MHz)

gNB MIMO

Channel model

gNB receiver noise figure (dB)

Calculated CSI effectiveness

Path-loss exponent

Shadow fading sigma (dB)

Delay spread (ns)

Mobility (km/h)

Coverage assumptions Calculated

Awaiting runRun the simulation to calculate path loss, RSRP, SINR, CQI and coverage radius.

Multi-Cell / Cluster Configuration Coverage topology

Number of sites

Sectors per site

Optimization clusters

Grouping method

Cell inventory CSV

Map/cluster status

Cluster morphology automatically follows the main SIM Scenario selection. Generate a synthetic multi-cell layout or import cell_id, latitude, longitude and optional cluster_id.

Sites

Synthetic/imported

Cells/sectors

Optimization entities

Clusters

Visualization groups

KPI assignment

Not assigned

15-minute records

Defines the site/cell topology used by the current simulation and geographic Results map. Set 1 site / 1 sector / 1 cluster for a single-cell run. The common trained PPO is evaluated independently for each cell; coordinated multi-cell RL and handover state evolution remain future work.

TDD Capacity Layer 1 Mid-band

Center frequency (GHz)

DL TDD fraction

gNB MIMO

Channel model

Beamforming gain (dB)

Rank adaptation efficiency

CSI feedback delay (ms)

TDD Capacity Layer 2 Upper mid-band

Center frequency (GHz)

DL TDD fraction

gNB MIMO

Channel model

Beamforming gain (dB)

Rank adaptation efficiency

CSI feedback delay (ms)

Capacity-stage outputs Calculated per layer

Awaiting runRun the simulation to calculate component-carrier SINR, CQI, spectral efficiency and capacity.

Carrier Aggregation Up to 4 DL CCs

Carrier Aggregation

Additional FDD CCs beyond anchor CC1

Total FDD + TDD component carriers

UL Carrier Aggregation

CC1 is always the FDD coverage anchor. In FDD-only mode, the “additional FDD CCs” value counts carriers beyond CC1: 1 additional means 2 total CCs. CC2–CC4 are configured below. The UE is assumed capable of receiving every selected CC.

CC2 source

CC2 FDD band

CC2 DL center frequency (MHz)

CC2 paired UL frequency (MHz)

CC2 FDD DL bandwidth (MHz)

CC2 FDD UL bandwidth (MHz)

CC2 TDD bandwidth (MHz)

CC3 source

CC3 FDD band

CC3 DL center frequency (MHz)

CC3 paired UL frequency (MHz)

CC3 FDD DL bandwidth (MHz)

CC3 FDD UL bandwidth (MHz)

CC3 TDD bandwidth (MHz)

CC4 source

CC4 FDD band

CC4 DL center frequency (MHz)

CC4 paired UL frequency (MHz)

CC4 FDD DL bandwidth (MHz)

CC4 FDD UL bandwidth (MHz)

CC4 TDD bandwidth (MHz)

Cross-carrier scheduling

CA combining efficiency

CA summary Network configured

Configured CCs

—

DL CC

Selected CCs

—

Network configured

Aggregated DL BW

—

MHz

CA efficiency

—

Component-carrier assignment CC1–CC4

CC	Source layer	Frequency	Bandwidth	Duplex	Status
Run the simulation to populate CC1–CC4.

Base scheduler PF

Scheduler

PF alpha

PF beta

CQI weight

Cell-edge weight

AI optimization Tunes selected scheduler

AI engine — select one

RL policy

Policy execution

Simulation repeatability

AI-assisted CSI reporting

Decision interval

Optimization goal

Predictive traffic tuning

Throughput weight

Fairness weight

Cell-edge weight

Energy weight

Confidence threshold

AI engine status Engine ready

Selected engine

PPO Pretrained

Optimization implementation

PPO model

RANOpt-PPO-v17.7-beta

Embedded demonstration policy

Stage-1 policy

Available

Offline-pretrained representation

Inference

Ready

Browser execution

Random seed

Fixed 12345

Repeatability mode

Training episodes

250,000

Stage-1 reference

The browser loads a baseline policy representation and can optionally continue local contextual-bandit training. v17 stores compatible checkpoints in Chrome IndexedDB and supports validated synthetic KPI CSV input. This is development/simulation training, not carrier-trained operation.

PPO Training Baseline model loaded

Built-in PPO training data

Loading built-in 30-cell × 96-sample synthetic dataset…

Optional Carrier CSV

Sampling interval

Random seed

Training episodes

Checkpoint interval

Stochastic expansion

Dataset status

Built-in multi-cell synthetic data loads automatically. Carrier CSV is optional; a valid carrier upload automatically continues training from the persistent PPO checkpoint.

Import model

Training rows

Automatic morphology-expanded records

Local episodes

Persisted contextual-bandit episodes

Mean reward

—

Current training batch

Checkpoint

Not saved

Chrome IndexedDB

Convergence

Not trained

Reward stability indicator

v17 training workspace ready. Load a validated CSV to begin.

Beta v17.7 uses a built-in 30-cell × 96-sample synthetic dataset and automatically expands it across Suburban, Urban and Dense Urban morphology variants. Normal SIM runs are inference-only. Optional carrier CSV input automatically continues training of the same persistent PPO.

3GPP-aligned KPI target profile Balanced KPI constraints

Target profile

RL optimization mode

Target median DL (Mbps)

Target median UL (Mbps)

Target 5%-tile DL (Mbps)

Target 5%-tile UL (Mbps)

Target fairness

Maximum PRB utilization (%)

UL signaling reduction target (%)

Targets are configurable engineering profiles used by Balanced KPI Optimization. The Rel-19 / emerging 6G profile is a research target profile, not a finalized 6G normative requirement.

No simulation results available

Configure the network parameters and press Run Simulation to generate new KPI results.

Simulation KPIs Awaiting run

Median DL UE rate

—

Mbps

Median UL UE rate

—

Mbps

5th percentile DL

—

Mbps

5th percentile UL

—

Mbps

Peak DL

—

Gbps

Peak UL

—

Gbps

Cell DL capacity

—

Gbps

Cell UL capacity

—

Gbps

Estimated VoNR capacity

—

simultaneous users · FDD overlay

Coverage radius

—

Cell spectral efficiency

—

b/s/Hz

Jain fairness

—

index

PRB utilization

—

AI capacity gain

—

AI confidence

—

UL UE signaling reduction

—

Optimization reward score

—

/100

Selected DL bandwidth

—

MHz

Selected component carriers

—

KPI target attainment Awaiting run

KPI	Achieved	Status
Run the simulation to compare the selected AI result with the configured thresholds.

Run the simulation to compare available cell capacity with the capacity required by the configured per-UE targets.

UE throughput distribution DL / UL

Coverage / SINR Distance profile

DL capacity by component carrier CC1–CC4 and total

Cell resource utilization Normal and Expert

Conspicuous AI impact AI OFF baseline vs AI ON

Only AI-sensitive KPIs are shown. Bar heights are normalized to AI OFF = 100; exact values and percentage improvement are printed on the chart.

Geographic Output Map Awaiting SIM

Map layer

Comparison

PPO objective

Map status

Run the simulation to generate cell-wise output for the selected cluster.

Map extent automatically fits all cells participating in the current simulation. Coverage is model-predicted, not measured drive-test coverage. Cluster grouping is visual; PPO inference is evaluated cell-by-cell.

Per-carrier DL capacity Expert mode

Link adaptation diagnostics Expert mode

Scheduler and fairness diagnostics Expert mode

AI optimization diagnostics Expert mode

Hierarchical calculation trace Intermediate values

Run the simulation to populate the calculation trace.

Validation checks Regression and consistency tests

Ready. Quick, Detailed, and NLP tests run automatically and report implementation-specific results.

NLP validation has not been run.

RANOpt AI — Interactive AI-Assisted 5G/6G RAN Decision Platform Beta v17.7

RANOpt AI Voice Assistant

Voice command history

Deployment Network geometry

UE capability Constrains all layers

Coverage FDD layer Anchor layer

Coverage assumptions Calculated

Multi-Cell / Cluster Configuration Coverage topology

TDD Capacity Layer 1 Mid-band

TDD Capacity Layer 2 Upper mid-band

Capacity-stage outputs Calculated per layer

Carrier Aggregation Up to 4 DL CCs

CA summary Network configured

Component-carrier assignment CC1–CC4

Base scheduler PF

AI optimization Tunes selected scheduler

AI engine status Engine ready

PPO Training Baseline model loaded

3GPP-aligned KPI target profile Balanced KPI constraints

No simulation results available

Simulation KPIs Awaiting run

KPI target attainment Awaiting run

UE throughput distribution DL / UL

Coverage / SINR Distance profile

DL capacity by component carrier CC1–CC4 and total

Cell resource utilization Normal and Expert

Conspicuous AI impact AI OFF baseline vs AI ON

Geographic Output Map Awaiting SIM

Per-carrier DL capacity Expert mode

Link adaptation diagnostics Expert mode

Scheduler and fairness diagnostics Expert mode

AI optimization diagnostics Expert mode

Hierarchical calculation trace Intermediate values

Validation checks Regression and consistency tests