According to Boehm[1] and Blottner[2], we can adopt the following succinct descriptions:
Verification: "Solving the equations right."
Validation: "Solving the right equations."
Verification and Validation (V&V) are both necessary processes for demonstrating the predictive capability of a computational model. These two processes can be defined more specifically as[3]:
Verification: Process for ensuring that a numerical model accurately represents the conceptual description of the physical model and the related analytical solution. In other words, the verification process involves proving that the code within the mathematical model and the numerical algorithm are accurately producing the computational model as intended.
Validation: Assesses the degree to which the computational model accurately represents the physics of the real world being modeled in an intended application. The validation requires the comparison between simulation and experimental data, where the predictive capability is evaluated along with the physical reality while addressing the uncertainties arising from both experiments and computations.