Evolutionary Computing - Introduction to Soft Computing

What is Computing?

  • The use or operation of computers.
  • Goal-oriented activity using algorithms.
  • Includes designing hardware/software, processing information, scientific research, intelligent systems, and media creation.

Soft Computing vs. Hard Computing

  • Hard computing requires precise analytical models and computation time.
  • Soft computing is tolerant of imprecision, uncertainty, partial truth, and approximation; it mimics the human mind.

Hard Computing

  • Relies on precision, certainty, and rigor.
  • Struggles with recognition problems, mobile robot coordination, forecasting, and natural language reasoning.

Soft Computing

  • Tolerant of imprecision and uncertainty.
  • Aims for tractability, robustness, and low costs.
  • Real-world problems are often imprecise, making precision costly or impossible.
  • Employs Fuzzy Logic (FL), Artificial Neural Networks (ANN), Evolutionary Computation (EC), Swarm Intelligence, Machine Learning (ML) and Probabilistic Reasoning (PR).

Implications of Soft Computing

  • Techniques like ANN, EC, and FL are used complementarily.
  • Neuro-fuzzy systems are increasingly common in consumer and industrial products.
  • Learns from data and generalizes by approximating outputs from unseen inputs using previous learned inputs.

Applications of Soft Computing

  • Handwriting recognition, automotive systems, image processing, architecture, expert systems, data mining, power systems, and control systems.

Future of Soft Computing

  • Represents a paradigm shift, mirroring the human mind's ability to process imprecise information.

AI and Soft Computing

  • ANN: Learning and adaptation.
  • Fuzzy Set Theory: Knowledge representation via Fuzzy if-then RULE.
  • Genetic Algorithms: Systematic Random Search.

Overview of Techniques

  • Fuzzy Logic Models: Representing imprecise human knowledge.
  • Neural Networks: Inspired by biological nervous systems for pattern recognition, prediction, etc., composed of interconnected neurons organized in layers (input, hidden, output).
  • Genetic Algorithms: Inspired by genetics and evolution, using "survival of the fittest" to solve complex design optimization problems with discrete, continuous, and nonlinear functions.