OptaPlanner

OptaPlanner is an open-source Java library for constraint solving and planning. It provides a framework for modeling and solving optimization problems in business domains such as scheduling, rostering, vehicle routing, and resource allocation.

Modeling approach: Users define a domain model with planning entities and planning variables. A planning solution contains the full set of entities and their assignments, along with a Score that reflects how well the constraints are satisfied. Constraints are declared as hard and soft rules using Constraint Streams (a Java-based DSL) or the Drools DRL language. The scoring mechanism updates incrementally as the solver makes changes to the solution.

Solving methods: OptaPlanner relies on metaheuristic search to explore the space of feasible solutions. Typical algorithms include Tabu Search, Simulated Annealing, Late Acceptance, and various local-search hybrids. It is designed to find high-quality solutions for large, complex problems and supports multi-threaded solving to enable parallel search.

Features and integration: The library provides annotations such as @PlanningSolution, @PlanningEntity, and @PlanningVariable to define the problem model, along with tools for benchmarking and debugging. It is designed to integrate with Java applications and works well with other Drools-based tooling in the ecosystem.

Applications and ecosystem: OptaPlanner is used for domains such as workforce rostering and nurse scheduling, class and university timetabling, vehicle routing, production planning, and cloud or data-center resource optimization. It emphasizes a declarative approach to modeling and relies on flexible constraint definitions to produce practical, near-optimal plans.

Licensing and resources: The project is released under the Apache License 2.0 and maintains comprehensive documentation, tutorials, and example domains to support learning and adoption.