Algorithm Architecture·3 min read·September 2026

The Five Named Dietary Relaxations: What Happens When Plans Are Infeasible

How constraint satisfaction algorithms gracefully diagnose mathematical bottlenecks.

Every nutrition app faces a fundamental dilemma when a user enters mutually contradictory constraints: does the software pretend the math works, hallucinate an impossible meal plan, or honestly diagnose the conflict?

1. The Infeasible Parameter Set Paradox

Consider a user who demands 1,200 calories per day, 160g of protein, a 15g net carb ceiling, and a vegan exclusion. In human biology, 160g of pure plant protein requires at least 1,450 to 1,900 calories.

An unconstrained AI model will invent fake recipes with fabricated macros. A deterministic solver recognizes that the mathematical solution space is null (empty).

When a user requests 1,200 kcal and 160g vegan protein, the mathematical solution space is empty. An honest solver diagnoses the conflict instead of faking the numbers.

2. The Five Named Relaxations Priority Cascade

When constraints conflict, Repast applies an explicit, prioritized hierarchy: 1. Hard Allergen Invariant: Never relaxed under any circumstance. 2. Hard Carb Ceiling: Never relaxed — your 20g net limit remains absolute. 3. Minimum Protein Floor: Maintained to prevent lean mass catabolism. 4. Caloric Band Tolerance: Allowed to expand by ±5% to reconcile whole-ingredient portion steps. 5. Recipe Variety Penalty: Softened to allow reliable ingredient chaining when food choices are tightly restricted.

KEY TAKEAWAY

Mathematical honesty beats algorithmic hallucination. When your dietary targets conflict, your planner should diagnose the bottleneck with precision.

HONEST CONSTRAINT SOLVER

Experience meal planning that never cheats the math.

Repast solves deterministic equations without hallucinating. If your parameters conflict, it provides precise diagnostic explanations.

Run honest meal planning on iPhone