Convergence Encyclopedia: C15 — Optimization Under Constraint / Pareto Fronts

F1 — Tier. T0 (Pareto optimality — mathematical definition) / T1 (ubiquitous instantiation in economics, biology, engineering, AI).

F2 — Sources.

• Pareto, V. (1906). Manuale di economia politica con una introduzione alla scienza sociale. Societa Editrice Libraria. (Pareto optimality: no individual can be made better off without making another worse off.)
• Dantzig, G.B. (1963). Linear Programming and Extensions. Princeton University Press. (Origins 1947.)
• Levins, R. (1966). “The strategy of model building in population biology.” American Scientist, 54(4), 421–431. (Evolutionary trade-offs.)
• Shoval, O. et al. (2012). “Evolutionary trade-offs, Pareto optimality, and the geometry of phenotype space.” Science, 336(6085), 1157–1160.
• Sutherland, W.J. (2005). “The best solution.” Nature, 435(7045), 569. (Review of optimization in biology.)
• Thermodynamic bounds: Seifert, U. (2012). “Stochastic thermodynamics, fluctuation theorems and molecular machines.” Reports on Progress in Physics, 75(12), 126001.

F3 — Domains. Economics (Pareto efficiency), biology (evolutionary trade-offs — e.g., growth vs. defense), engineering (multi-objective optimization), AI (multi-objective reinforcement learning), thermodynamics (entropy production bounds).

F4 — Scale. Molecular motors (~10⁻⁹ m) → economic systems (~10⁹ m, global).

F5 — Falsifier. A real system (biological, economic, or engineered) that is Pareto-dominated on all relevant objectives by an alternative that is actually reachable — i.e., a system that persists despite being strictly worse than an available alternative on every dimension. (Note: persistent suboptimality is common; the falsifier requires suboptimality with a reachable superior alternative. The challenge is defining “reachable.” See rival below.)

F6 — Rival (strongest form). Pareto optimality is a static description, not a dynamic process. Real systems are rarely on the Pareto front; they are constrained by history, path dependence, and incomplete information. The appearance of trade-offs is a sign of constraint, not optimization. Shoval et al. (2012) demonstrated Pareto-like geometry in phenotype space, but this is consistent with constraint satisfaction, not active optimization. (Gould & Lewontin 1979 “spandrels” argument extended.)

F7 — Independence. HIGH. Pareto (economics, Lausanne), Dantzig (operations research, RAND/Berkeley), Levins (theoretical biology, Harvard), Seifert (statistical physics, Stuttgart) — four fields, no shared institutional lineage. The mathematical framework (multi-objective optimization) is shared, but the empirical discoveries of trade-offs were independent.

F8 — Pattern type. Mathematical.

F9 — Maps. A2 (thermodynamic/computational), A3 (pattern-dynamics).

---

Corpus map

• Previous: Convergence Encyclopedia: C14
• Next: Convergence Encyclopedia: C16
• Encyclopedia start: The Schema
• Same node, other planes: Catalogue node C15 · Catalogue hub
• Edges touching C15: convergence edge 2
• Kin corpora: Total Structure · Signature of the Grain