An Electromagnetic Attraction and Repulsion Simulated Techniques-based Optimalization Software Framework and Applied Solving Systems
| Autor: | Chin-Yuan Kang, 康志遠 |
|---|---|
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 93 The electromagnetic-like mechanism (EM) optimization algorithm is a new developing heuristic algorithm. Typical EM algorithm applies to solve continuous optimization problems without general constraints. This research analyzes the detail of EM algorithm and generalizes the compatibility of all kinds of optimization problems, and develops a meta level EM algorithm for solving all kinds of optimization problems. This research probes into the design pattern of EM and carries on systematic function analysis and structural design, then uses object-oriented technique to develop an electromagnetic attraction and repulsion simulated techniques-based optimalization software framework and applied solving systems. This research has developed four EM-based optimization system. These software systems include an MS Excel interfaced optimization package, EMBOS-E (EM-Based Optimization System for Excel); a C# programming languages interfaced software package, EMBOS-NET (EM-Based Optimization System for .NET); a software development kit for creating EM optimization solvers, EMBOSDK (EM-Based Optimization Software Development Kit); and an object-oriented software framework for optimization system developments, EMBOF (EM-Based Optimization Framework). Naive users can use EMBOS-E to setup and run their optimization problems on the Excel platform. Intermediate users can use EMBOS-NET and C# programming language to define their problems and implement additional procedures. Their written code will be runtime compiled, linked, and executed. Advanced users can either use EMBOSDK or EMBOF to develop their own solving packages. EMBOSDK uses API function calls to construct their systems. Object-oriented techniques are used to build optimization systems by using classes of EMBOF. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|