Skip to content

Based on Pigeon-Inspired Optimization (PIO), this paper proposes the Adaptive Learning Pigeon-Inspired Optimization and Opposition-based Learining with Sine and Cosine (ALPIO-OBL-SC) to solve the inherent defects such as poor parameter accuracy, and improve the adaptive ability of the FPN model.

Notifications You must be signed in to change notification settings

chen0495/ALPIO-OBL-SC

Repository files navigation

本文基于改进鸽群算法对模糊Petri网参数进行优化,详细介绍请到Relaese下载论文PDF
实验主体文件为“PIO_反向学习_正余弦.ipynb”,请在对应环境下运行该文件

介绍

模糊Petri网(FPN)基于专家知识库模糊产生式规则对模糊或不精确的知识进行表述。但FPN模型中的参数(如权值、阈值和确信度)依赖专家知识和经验,因而存在参数精度难以保证、自适应能力较差、泛化能力弱等缺陷。而群智能优化算法则具有较强的自适应能力,能够基于一组初始参数进行参数自优化,因此将改进鸽群算法引入FPN的参数优化问题当中可以弥补FPN的上述缺陷,提高其推理结果的准确度,使FPN的性能在专家知识库等应用领域更加高效可靠。
本文提出一种基于反向学习策略和正弦余弦优化策略的改进鸽群算法(Adaptive Learning Pigeon-Inspired Optimization and Opposition-based Learining with Sine and Cosine,ALPIO-OBL-SC)对FPN参数进行优化。

实验结果展示

实验基于此FPN模型:
FPN模型
测试函数收敛性测试对比结果:
收敛性测试对比
FPN收敛性测试对比结果:

实验环境

实验环境:

  • Intel(R) Core(TM) i7-8550U CPU@1.80GHz
  • win10 Pro操作系统
  • 16G内存
  • Python3.8
  • Jupyter Notebook 6.4.10

目录结构

  • PIO鸽群优化算法
    • Function.py 测试函数包
    • main.py 标准鸽群算法入口,已废弃
    • PIO.py 标准鸽群算法主体,已废弃
    • PIO_WOA.ipynb PIO、ALPIO-OBL-SC、WOA三类算法对比展示,jupyter notebook
    • PIO_反向学习_正余弦.ipynb PIO、ALPIO-OBL-SC两类算法对比展示,jupyter notebook
    • ReadMe.txt 说明文件.txt
    • README.md 说明文件.md
    • 改进鲸鱼算法.ipynb WOA算法展示

About

Based on Pigeon-Inspired Optimization (PIO), this paper proposes the Adaptive Learning Pigeon-Inspired Optimization and Opposition-based Learining with Sine and Cosine (ALPIO-OBL-SC) to solve the inherent defects such as poor parameter accuracy, and improve the adaptive ability of the FPN model.

Resources

Stars

Watchers

Forks

Packages

No packages published