Recent Developments in Pyomo

Computational tools for modeling mathematical programs are widely used within both academia and industry. Available commercial and open-source modeling packages support generic modeling by separating modeling constructs from instance data through concepts like sets, parameters, and parameterized constraints. However, their model representations are limited to constructs that directly correspond to established solver inputs. In general, this implies that mathematical programs must be expressed as either linear or nonlinear algebraic models; that is, a list of variables, an algebraic objective expression, and a list of algebraic constraints.

In contrast, the Pyomo environment is designed to be an open Python-based environment for developing and exploring complex, optimization-based analytical strategies. This includes support for extending the modeling environment to include non-algebraic modeling constructs, developing automated model transformation routines, and automating problem decomposition and interrogation.

In this presentation we highlight recent developments in the Pyomo project, including new language constructs for supporting modeling with units, preliminary support for native automatic differentiation, and new strategies for accelerating the solution of block-decomposable models.