Abstract
Several interesting libraries for optimization have been proposed. Some focus on individual optimization algorithms, or limited sets of them, and others focus on limited sets of problems. Frequently, the implementation of one of them does not precisely follow the formal definition, and they are difficult to personalize and compare. This makes it difficult to perform comparative studies and propose novel approaches. In this paper, we propose to solve these issues with the General Purpose Optimization Library (GPOL): a flexible and efficient multipurpose optimization library that covers a wide range of stochastic iterative search algorithms, through which flexible and modular implementation can allow for solving many different problem types from the fields of continuous and combinatorial optimization and supervised machine learning problem solving. Moreover, the library supports full-batch and mini-batch learning and allows carrying out computations on a CPU or GPU. The package is distributed under an MIT license. Source code, installation instructions, demos and tutorials are publicly available in our code hosting platform (the reference is provided in the Introduction).
Highlights
Thinking about something in abstract terms involves considering its central meaning or identifying overarching themes and fundamental issues that can apply across contexts
We propose to solve these issues with the General Purpose Optimization Library (GPOL): a flexible and efficient multipurpose optimization library that covers a wide range of stochastic iterative search algorithms, through which flexible and modular implementation can allow for solving many different problem types from the fields of continuous and combinatorial optimization and supervised machine learning problem solving
As noted from the abovementioned algorithms, in this release, we focus our attention on stochastic iterative-search algorithms, in which the search procedure, and the outcome involve some degree of randomness
Summary
Thinking about something (an object or an event) in abstract terms involves considering its central meaning or identifying overarching themes and fundamental issues that can apply across contexts. The first release of GPOL, presented in this document, implements random search, hill climbing, simulated annealing, genetic algorithm, geometric semantic genetic programming, differential evolution (represented through its numerous mutation strategies), and particle swarm optimization (present in its synchronous and asynchronous variants). Modular implementations of random search, hill climbing, simulated annealing, and the genetic algorithm allow one to adapt them to potentially any type of solution representation, as such problem types From this perspective, the random search unfolds into tree-based random search. The library supports the application of algorithms on several different types of problems, including continuous and combinatorial problems, and supervised machine learning (approached from the perspective of inductive programming and exploiting batch-training) It offers a broad range of operators, including some of the most recent developments in the scientific community, such as geometric semantic operators and evolutionary demes despeciation algorithm. The GitLab’s repository https://gitlab.com/ibakurov/general-purpose-optimization-library (accessed on 15 May 2021) contains the sourcecode, installation guidelines, additional demonstrations and benchmark scripts, namely end-to-end integration tests in the folder “main” and detailed tutorials in the folder “tutorials_jupyter” provided as Jupyter notebooks
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
