Symbolic regression that aims to detect underlying data-driven models hasbecome increasingly important for industrial data analysis. For most existingalgorithms such as genetic programming (GP), the convergence speed might be tooslow for large-scale problems with a large number of variables. This situationmay become even worse with increasing problem size. The aforementioneddifficulty makes symbolic regression limited in practical applications.Fortunately, in many engineering problems, the independent variables in targetmodels are separable or partially separable. This feature inspires us todevelop a new approach, block building programming (BBP). BBP divides theoriginal target function into several blocks, and further into factors. Thefactors are then modeled by an optimization engine (e.g. GP). Under suchcircumstances, BBP can make large reductions to the search space. The partitionof separability is based on a special method, block and factor detection. Twodifferent optimization engines are applied to test the performance of BBP on aset of symbolic regression problems. Numerical results show that BBP has a goodcapability of structure and coefficient optimization with high computationalefficiency.
展开▼
