A genetic algorithm based welding planner capable of using parametric features to determine the manufacturing cost of welded assemblies has been developed. Parametric weld feature information obtained from a CAD system is translated to a hybrid traveling salesman problem. A genetic algorithm is used to search for a sequence of welds that minimizes the cost of the welded assembly. Emphasis is placed on developing a mechanistic, feature based DFM tool, with the aim of rapidly providing welded assembly cost feedback to the designer in a CAD environment, while maintaining a balance between accuracy and computational speed.
A Genetic Algorithm and Support Vector Regression based Hybrid Cost Estimation Model for Feature-based Plastic Injection Products
