Product-Mix optimization through soft computing

In this paper the Product Mix optimization is considered using the Takagi-Sugeno Method. The fuzzy inference system, implemented in the frame work of adaptive networks, is through soft computing. The system uses an architecture called Adaptive-Network-based Fuzzy Inference System (ANFIS), which generates a set of fuzzy if-then rules (human knowledge or natural language) with an appropriate membership function and stipulated input-output data pairs. It is a fusion of Neural Network and Fuzzy Logic in a Neuro-Fuzzy model that has the ability for learning and read if-then rules at the same time. Supply chain parameters such as, demand, resource-availability, buffer-size etc. are related to the systems performance index such as yield, profit, customer satisfaction of the company. Using the linear and/or non-linear regression model the dynamics of the supply chain based on the above parameters are studied at the production and enterprise level. Since the approach is data centric, different set of data can be used for training, testing and validation, from which the system can learn and help managers in the decision making. The system implements the genfis and anfis functions given in MATLAB^reg and the results so obtained, demonstrates the use of neuro-fuzzy model, for a product mix, in the decisionmaking in supply chain.