Скачать или смотреть DecisionCAMP-2021: "A modern OO/FP constraint solver" by Geoffrey De Smet (Red Hat)

Traditional constraint solvers don't fit well with Object Oriented Programming (OOP) or Functional Programming (FP). An OO/FP constraint solver is different: Users input data in their OO domain model, instantiated from their own classes, using annotations to define optimization variables. They are not limited to numeric variables with symptoms of numerical instability. Users write their constraints in a OO/FP code, reusing existing code. They don’t need to translate them into mathematical equations with linear or quadratic limitations. And they can unit test them. In this presentation I‘ll compare traditional solvers to an OO/FP constraint solver like OptaPlanner.
Keywords: artificial intelligence, operations research, constraint solver, java, open source