Syllabus Soft Computing - (BE05043021) Total Credits = TH / 30 Assessment Pattern and Marks Total Marks Theory Tutorial / Practical ESE (E) PA (M) PA (I) PBL (I) ESE (V) 03 70 30 20 30 50 200 Unit No. Content 1 Introduction to Soft Computing Hard Computing : features and examples; Soft Computing : features and examples; Hybrid Computing and its significance; distinction between hard and soft computing; components of soft computing; role of uncertainty and imprecision in real-world problem solving. (Chapter - 1) 2 Optimization Introduction to optimization; classification of optimization problems; principle of optimization; duality principle; traditional methods : Exhaustive Search Method, Random Walk Method and Steepest Descent Method; drawbacks of traditional methods; overview of Simulated Annealing and Particle Swarm Optimization; basic comparison of PSO and GA. (Chapter - 2) 3 Genetic Algorithms Working cycle of Genetic Algorithm; binary-coded GA; selection, crossover and mutation; schema theorem; limitations of binary-coded GA; GA parameter setting; penalty function approach for constraint handling; advantages and disadvantages of GA; real-coded GA with BLX-alpha and SBX crossover; mutation operators; micro-GA. (Chapter - 3) 4 Introduction to Fuzzy Sets Crisp sets : notation, operations and properties; fuzzy sets : representation and difference from crisp sets; support, height, alpha-cut, normality and convexity; fuzzy set operations; fuzzy relations and composition; properties of fuzzy sets; measures of fuzziness and inaccuracy. (Chapter - 4) 5 Neural Networks Biological neuron and artificial neuron; layer and multi-layer architectures; static versus dynamic neural networks; supervised, unsupervised, incremental and batch training; Multi-Layer Feed-Forward Neural Network (MLFFNN) : forward calculation, back-propagation learning, design steps, advantages and limitations; simple numerical understanding. (Chapter - 5) 6 Hybrid Soft Computing Fuzzy-Genetic Algorithm : concept, structure and working principle; Neuro-Fuzzy System based on Mamdani approach; basic tuning using back-propagation and Genetic Algorithm. (Chapter - 6) 7 Applications of Soft Computing Applications in intelligent autonomous robots, environment information collection, motion planning and motor control; applications in data analysis, classification and predictive modelling; illustrative case studies in pattern recognition, optimization, diagnosis and engineering decision-making. (Chapter - 7)
