The evolutionary population synthesis (EPS) is the technique to model the spectrophotometric properties of stellar populations, that uses the knowledge of stellar evolution. This approach was pioneered by B. Tinsley (see Section 3.1) in a series of fundamental papers, that provide the basic concepts still used in present-day models. The target of EPS models are those stellar systems that cannot be resolved into single stars, like galaxies and extra-galactic globular clusters. The comparison with the models aims at providing clues on the ages and element abundances of these unresolved stellar populations, in order to constrain their formation processes, and finds ubiquitous astrophysical uses. The simplest flavour of an EPS model, called Simple Stellar Population (hereafter SSP), assumes that all stars are coeval and share the same chemical composition. The advantage of dealing with SSPs is twofold. First, SSPs can be compared directly with globular cluster (hereafter GC) data, since these are the "simplest" stellar populations in nature. This offers the advantage of calibrating the SSPs with those GCs for which ages and element abundances are independently known ([35]). This step is crucial to fix the parameters that are used to describe that part of the model "input physics" that cannot be derived from first principle (convection, mass loss, mixing). The calibrated models can be applied with more confidence to the study of extragalactic stellar population. Second, complex stellar systems which are made up by various stellar generations are modelled by convolving SSPs with the adopted star formation history (e.g. [1], [36], [47], [19], [2]), therefore the deep knowledge of the building blocks of complex models is very important.
The article starts with a description of stellar population models in terms of ingredients, assumptions and computational technique (Section 2), which is followed by a historical overview of the evolution of these models until the most recent results (Section 3).