A Criteria Transformation Approach to Pattern Matching based on Non-Linear Parameter Optimization

摘要

This paper presents a concept for pattern matching based on a parameter optimization system for approximative numerical calculation of some parameter combination under soft and hard constraints. The concept uses a non-linear parameter optimization method with an iterative variation of parameters. The paper focuses on the information modeling process to migrate problem-domain specific criteria into optimization-compatible objects suitable for a standardized parameter optimization procedure. A step-by-step transformation process is presented and implemented in object-oriented programming: classes and interfaces. The method is applicable to a wide range of pattern-matching problems due to its flexibility and extensibility.