Efficient Optimization through Normalized Variable Scaling and Selective Term Pruning

This paper presents an innovative pruning algorithm designed to enhance computational efficiency of optimization-related tasks by reducing the number of computations. The algorithm, which is based on optimization variable normalization and subsequent term pruning according to a set relative threshold, offers balance between accuracy and computational complexity of the solution, making it suitable for performance-critical applications. The trade-off between accuracy and complexity in this approach is expressed by metrics of Pareto efficiency and relative efficiency.