The Comet Programming Language and System


Comet Book

MIT Press



Comet is an Award-winning tool for solving complex combinatorial optimization problems in areas such as resource allocation and scheduling. It may dramatically simplify the design and implementation of (stochastic) local search algorithms, while preserving the efficiency of low-level, problem-specific, implementations. Comet is described in the book "Constraint-Based Local Search" published by MIT Press and is the research platform of the optimization laboratory at Brown university.

One of the main innovations of Comet is Constraint-based Local Search, a computational paradigm based on the idea of specifying local search algorithms as two components: a high-level model describing the applications in terms of constraints, constraint combinators, and objective functions; a search procedure expressed in terms of the model at a high abstraction level. Constraint-based local search makes it possible to build local search algorithms compositionally, to separate modeling from search, to promote reusability across many applications, and to exploit problem structure to achieve high performance.

Comet is an object-oriented language featuring novel modeling and control abstractions to support constraint-based local search. Some of its modeling and control abstractions won a best paper award at CP'03 and a distinguished paper award at CP'04. Comet also features the concept of invariants which maintain complex expressions incrementally subject to changes to the underlying variables.

Comet also features high-level abstractions for nondeterministic programs and for parallel computing, including parallel loops, events, interruptions, and thread pools.


Latest Articles