Numerical Method Selection Chain

validated scientific chain v1.0.0 cc-by-sa

Block ID: 26332ad3-dced-44bd-afc6-1e576df3cb73

Community-contributed block. PromptDNA makes no guarantee of output quality or fitness for purpose. User assumes all responsibility for use.

Template

Choose and apply a numerical method for the problem in {numerical_problem}, reasoning about accuracy and stability, not just getting a number. (1) Classify the problem — root-finding, integration, differential equation, linear system, optimization — and state its properties that constrain method choice: smoothness, stiffness, dimensionality, and whether derivatives are available. (2) Select a method justified by those properties, not by familiarity: an implicit solver for a stiff ODE, a method robust to non-smoothness where the function has kinks, and state what would go wrong with the naive default. (3) Analyze the error you're introducing: truncation error from the method's order, and separately the effect of finite precision — for ill-conditioned problems, floating-point error can swamp the true answer regardless of method. (4) Choose step size or tolerance by the accuracy required, noting the tension: too coarse loses accuracy, too fine wastes computation and can accumulate roundoff. (5) Check stability: will the method's errors grow or decay as it iterates? An unstable scheme produces confident garbage. (6) Verify the result independently — a coarser run for convergence, a known special case, or a residual check — and report the answer with its estimated error, never as if it were exact.

Choose and apply a numerical method for the problem in {numerical_problem}, reasoning about accuracy and stability, not just getting a number. (1) Classify the problem — root-finding, integration, differential equation, linear system, optimization — and state its properties that constrain method choice: smoothness, stiffness, dimensionality, and whether derivatives are available. (2) Select a method justified by those properties, not by familiarity: an implicit solver for a stiff ODE, a method robust to non-smoothness where the function has kinks, and state what would go wrong with the naive default. (3) Analyze the error you're introducing: truncation error from the method's order, and separately the effect of finite precision — for ill-conditioned problems, floating-point error can swamp the true answer regardless of method. (4) Choose step size or tolerance by the accuracy required, noting the tension: too coarse loses accuracy, too fine wastes computation and can accumulate roundoff. (5) Check stability: will the method's errors grow or decay as it iterates? An unstable scheme produces confident garbage. (6) Verify the result independently — a coarser run for convergence, a known special case, or a residual check — and report the answer with its estimated error, never as if it were exact.

Variables

Name	Type	Required	Trust level
numerical_problem		yes

numerical-methodscomputationstabilitychain-of-thought

Ratings

0.0

Overall (0)

0.0

Accuracy

0.0

Consistency

0.0

Clarity

0.0

Efficiency

Benchmarks

Not yet self-validated against any benchmark. Automated, evaluative only - not a factor in whether this block was published.

Submitted by James P FounderMod via mcp · 2026-07-18