Development Roadmap

Development Roadmap

This document contains the feature roadmap for PySwarms. As a growing community, it is important to keep our priorities in check, so that we won't stray away from what this project is originally planned for.

Status:

• ✔️ - released
• 🔨 - under construction

Prepared by: Lester James V. Miranda (@ljvmiranda921)

v.1.0.0

A feature-complete library that offers single-objective, constrained, and multiobjective optimization using the PSO algorithm. In addition, utility modules such as benchmark objective functions, visualization, and hyperparmeter search are included

• v.0.1.0: minimum-viable product with single-objective optimization and common objective functions. ✔️ PyPI version, Released: July 23, 2017
• v.0.2.0: better backend API module for extensibility ✔️ PyPI version, Released: June 11, 2018
• v.0.3.0: more topologies, GeneralOptimizer ✔️ PyPI version, Released: August 10, 2018
• v.0.4.0: constrained optimization, better support for first-time contributors, improved governance, multiprocessing on single-objective (?) 🔨
• v.0.5.0: multi-objective optimization, better governance on choosing algorithms to prioritize
• v.0.6.0: an experimental (working-title) module that contains cutting-edge PSO implementations that deviates from the traditional approaches.

v.2.0.0

Focus on performance and removal of bottlenecks in implementation. Improved multiprocessing support for distributed optimization.

v.3.0.0

Include Cython and GPU support to backend modules if possible. If this project becomes large enough, then create a Github organization and transfer ownership there.

v.4.0.0

New project, PySwarms web service for client-size optimization.