Hybrid Systems in Soft Computing

Hybrid Systems computing uses more than one computational technique to solve various real world problems. This integration of multiple systems in one enables us to get highly intelligent results. These results are potent as well as adaptive to any new environment.

Most of the techniques are inspired by natural intelligent biological processes. One of the most popular examples of Hybrid System is Neural Networks. Other famous examples include Fuzzy Systems and Genetic Algorithms. There are the various types of Hybrid Systems:

Sequential Hybrid Systems

  • As the name suggests, this system exhibits a pipe-like flow structure
  • It is also the weakest hybrid system
  • Sequential Hybrid System is characterized by a very straightforward structure and exhibits integration or congregation of a variety of technologies.
  • GA preprocessor is an ideal example of these types of systems.
  • You can understand the basic functionality of a Sequential Hybrid System in the following image.
Sequential Hybrid technology

Embedded Hybrid Systems

  • Embedded hybrid systems undertake an intertwined process flow in soft computing.
  • In this type of system no technology can be used without the involvement of other hybrid technologies.
  • Most popular example is the Neural Network and Fuzzy Logic hybrid system. 
  • An Embedded Hybrid System’s basic functionality is displayed in the following image
Embedded System

Auxiliary Hybrid System

  • In this system a particular technologies includes the other technology as a subroutine. This subroutine technology can called again to process information that will be used further in the overall system. 
  • Examples include neuro-genetic system which uses Genetic algorithms to optimize the neural networks input parameters
  • An Auxiliary Hybrid System works like this
Auxiliary Hybrid System

Applications of Hybrid System

Hybrid Systems are frequently used for Medical Image classification and Process control. Some others applications are:

  • Solving various real life problems and predictive outcomes.
  • Areas such as aviation, business development etc.
  • Solving real world problems that requires human like intelligence for decision making.
  • Process Control for various industries.

Examples of Hybrid Systems Explained

Neuro-Fuzzy Hybrid system

This system determines its input parameters by processing data with the help of machine learning taken from neural network/neural engine. These types of systems are also useful in applications such as mapping as well as carrying out various other complex processes.

It combines the learning of neural networks with the reasoning of fuzzy systems which enables it to make decisions best done with human intelligence. It enhances the flexibility and improves decision making so that it is able to adapt to any uncertain or new environment.

Advantages

  • User friendly approach
  • Easy to understand

Disadvantages

  • Difficult to interlink this model with classic systems
  • Limited work is available for study

Genetic-Neuro System

The Genetic-Neuro Fuzzy System is another successful hybrid system that uses parameters of Neural Networks and properties of genetic algorithm. In this type of system, the learning process of neural networks is done by using genetic algorithms. Many Chromosomes of the possible Neural Network is generated. Some advantages and disadvantages are given below:

Advantages

  • Enables best results due to genetic algorithm 
  • It is very robust

Disadvantages

  • Inputs must be changed every time before loading into the system for processing
  • Not very easy to understand and use

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *