VIII.10 / QUANTUM FOUNDATIONS
Computability, predictability, and physical realizability
What resources would a physical system require to implement the computation specified on paper?
Start with the question.
Computability, predictability, and physical realizability ask different questions about a model. Turing's formal analysis addresses which procedures a machine can carry out in principle, while Lorenz's dynamical example concerns sensitivity in a specified evolving system. A finite calculation can be computable yet expensive or sensitive to imperfect inputs. A mathematical construction also needs a physical implementation before its resources describe an actual device.
Separate three reasons a forecast can fail.
Consider a simulation whose update rule is known. Compare failure caused by insufficient runtime, uncertain initial conditions, and an incorrect physical model. Keep separate records for convergence, sensitivity, and agreement with observations. Increasing computational power may help the first problem while leaving the other two substantially unchanged.
What would you inspect?
- Algorithmic resource requirements, input precision, and numerical convergence.
- Agreement between the mathematical model and independently measured behavior.
Take the next step.
Classify one prediction limit by its actual cause and identify the measurement, algorithmic change, or model revision that could improve it.
Read the supporting material.
- On Computable Numbers, with an Application to the Entscheidungsproblem
Alan M. Turing / Stanford-hosted original paper · Formal computation and decision limits; does not establish practical prediction limits for every finite experiment.
- Deterministic Nonperiodic Flow
Edward N. Lorenz / MIT-hosted original paper · A deterministic dynamical-system example and prediction sensitivity.
