ATLASResearch
methods
Quantitative/ Formal analysis

Computational complexity analysis

Reason about resource growth as inputs scale

Complexity analysis derives how time, space or other resources grow with input size under a specified computation model. Worst-case, average-case, expected and amortised bounds answer different questions. Big O supplies an upper bound, while Theta states a tight asymptotic order. Mathematical analysis complements implementation benchmarks by separating scaling claims from hardware-specific timings.

WHEN IT FITS

Choose it when the question concerns scalability, algorithmic feasibility or guarantees over classes of inputs, and the algorithm, cost model and relevant input size parameters can be defined precisely.

Strengths

  • Provides guarantees beyond tested input sizes
  • Separates algorithmic growth from a particular machine

Limitations

  • Asymptotic bounds can hide substantial constants
  • Average-case claims depend on an input distribution

Know the boundary

Fast benchmark timings do not prove a complexity bound, and a better asymptotic bound need not be faster at small sizes.

USED ACROSS
Computer science