
Renormalization is a method in quantum field theory that removes infinities from calculations by adjusting parameters to match physical observations.Assembling a robot from a construction set with defective parts—each part is slightly larger or smaller than the specified size. If you just follow the instructions, the robot will fall apart. But the master knows which washers to add or how much to file to make everything work. Renormalization is the "adjustment" of infinite quantities in calculations so that the theory predicts reality.
The birth of renormalization is associated with H. Bethe's work on the Lamb shift (1947), followed by the covariant formulation of J. Schwinger, S. Tomonaga, and R. Feynman (1948), unified by F. Dyson. In 1954–55, M. Gell-Mann and F. Low introduced the concept of the renormalization group. G. 't Hooft and M. Veltman in 1971 proved the renormalizability of spontaneously broken gauge theories. K. Wilson in the 1970s developed the renormalization group in parameter space, clarifying the connection to critical phenomena and effective field theories, earning him the Nobel Prize in 1982.
The initial Lagrangian of quantum electrodynamics contains "bare" parameters e0, m0, ψ0. At order e0^2, vacuum polarization diagrams yield a logarithmically divergent integral. Regularization is introduced—for example, a cutoff momentum Λ or dimensional regularization d=4−2ε, after which the integral becomes finite. Then the parameters are split: e0 = eR + δe, m0 = mR + δm, ψ0 = √Zψ ψR, where the counterterms δe, δm, Z are chosen to cancel the divergences as Λ→∞ (or ε→0) in each loop. The renormalization group equation: μ dg/dμ = β(g) sets the dependence of the renormalized constant g on the scale μ. In QED, β(e) = e^3/(12π^2) + …; the positive sign means e grows with energy. For non-Abelian fields (QCD), the first coefficient is negative, causing asymptotic freedom. The effects of quantum fluctuations enter the observed charge: α(μ^2) = α(μ0^2) / [1 − (α(μ0^2)/(3π)) ln(μ^2/μ0^2)] in QED. A key principle is separation of scales: physics at large distances does not depend on the details of ultraviolet cutoffs. The method is closely related to the concepts of quantum field and Feynman diagrams.
Physicists long disliked renormalization, considering it cheating. Dirac said it was "not sufficiently beautiful mathematics," but without it, one could not explain why electrons have precise mass and charge.
| β_ρ | variable | beta function for ρ joule per cubic metre |
| M | variable | renormalization scale — energy joule |
| ρ | variable | running parameter joule per cubic metre |
| ∂/∂M | operator | partial derivative with respect to scale M — partial derivative |
| m_phys | variable | physical (renormalized) mass — mass kilogram |
| m_0 | variable | bare (Lagrangian) mass — mass kilogram |
| Z | variable | field strength renormalization constant |








