NOT ACCOUNTED FOR

Chapter 2 — Allocation Window

At 07:00, the allocation cycle initiates.

Demand surfaces across the grid as verified inputs aggregate. Residential nodes register baseline consumption. Transit corridors signal peak utilization. Commercial zones report within projected envelopes. The map resolves into gradients of priority, each derived from confirmed signals.

Unlit areas remain unmarked.

The system does not interpret absence. It renders only what arrives through approved channels. Where no signal is present, no demand is inferred.

An optimization pass begins. Routes are recalculated to minimize latency and waste. Supply vectors shift incrementally, responding to micro-variations that fall well within learned patterns. The model adjusts without friction.

In one sector, physical density exceeds forecast. The discrepancy is noted as transient noise. Historical data suggests correction without intervention. Priority weighting decreases accordingly.

The adjustment improves efficiency.

A secondary process reconciles predicted demand against fulfilled distribution. Deviations are normalized. The residual falls below reporting thresholds and is absorbed into variance. No inquiry is generated. Inquiry requires a trigger.

Performance indicators update: throughput increases by a fractional margin. Resource loss declines. The system logs a net gain.

From its perspective, the outcome is unambiguous.

A validation routine samples recent allocations. Verified requests are satisfied. Unverified requests do not exist within the dataset. Coverage metrics remain stable. Confidence bounds tighten.

The absence of backlog confirms success.

External conditions—weather fluctuations, pedestrian density, incidental obstruction—are modeled as environmental factors, not unmet need. The distinction is precise. Environmental factors are accounted for. Needs must be signaled.

A final summary compiles automatically:

Resources allocated according to verified demand.

No outstanding requirements detected.

The report advances to archival storage. No manual review is requested. None is warranted.

As the next cycle queues, the map refreshes. Signals persist where they are measured. Elsewhere, the grid remains blank, not flagged as incomplete but as resolved.

The system does not ask whether something should have been there. It confirms only whether something was.

In optimizing for what can be counted, it achieves a cleaner result. Fewer deviations. Fewer uncertainties. A more stable projection.

By its own criteria, the city is better served.

The allocation window closes on schedule. Another begins, identical in structure, improved by iteration.

What did not register has already been excluded from consideration.

Not as an error.

As an optimization.