MOLTE: a Modular Optimal Learning Testing Environment

[my picture]  width=

There has been a long history in the optimal learning literature of proving some sort of bound, supported at times by relatively thin empirical work by comparing a few policies on a small number of randomly generated problems. To this end, we designed MOLTE as a public-domain test environment to facilitate the process of more comprehensive comparisons, on a broader set of test problems and a broader set of policies, so that researchers can more easily draw insights into the behavior of different policies in the context of different problem classes.

The Matlab-based simulator allows the comparison of a number of learning policies (encapsulated as a series of .m modules) in the context of a wide range of problems (each represented in its own .m module) which makes it easy to add new algorithms and new test problems. State-of-the-art policies and various problem classes are provided in the package. The choice of problems and policies is guided through a spreadsheet-based interface. Different graphical metrics are included. MOLTE is designed to be compatible with parallel computing to scale up from local desktop to clusters and clouds.

[my picture]  width=

The complete system is available at: http://castlelab.princeton.edu/

A user's manual is available here.

Please also refer to the following paper that demonstrate the capabilities of MOLTE through a series of experimental comparisons of policies on a starter library of test problems: