----------------------------------------------------------------------------------------------------

Innovation has always been a tangled topic , Most graduate projects need algorithm innovation , And doctoral students need a lot of innovation to graduate . here , Based on the team's work experience in recent years , How to carry out reasonable algorithm innovation .

one , Innovation Perspective

usually , We use an algorithm , Here's a simple particle ,PSO Particle swarm optimization algorithm , We use simulation , The convergence speed of the algorithm will be obtained , Simulation accuracy and other parameters . If we need to innovate the algorithm , Generally, the performance index of the original algorithm should be considered , For example, the improvement of convergence speed and accuracy , For some low requirements , Under the condition that the convergence rate remains unchanged , Improve accuracy , Or with the same accuracy , Faster convergence , If the requirements are high , At the same time, how to improve the convergence speed and accuracy . Normally , in view of this situation , We need to consider other similar algorithms , These algorithms must have some characteristics of fast convergence and high accuracy , In this way, we can integrate the advantages of various algorithms , So as to realize the innovation of the algorithm and improve the performance of the original algorithm .

For low-level innovation , This is the general idea .

two , Innovation basis

What is the basis for innovation ? The so-called innovation basis , Is the innovative algorithm we use , The angle considered has theoretical basis , We can't modify the original algorithm out of thin air , In this way, even if a better solution of the layout is obtained , But it can not prove that the overall situation is better . therefore , As I said , We need to select some other off the shelf algorithms with better performance than the original algorithm , Integration of algorithms .

three , Verification of innovation

For the improved algorithm , We need to use a large number of test samples for comparative analysis , Verify whether each performance index of the algorithm is improved compared with the original algorithm .

four , Complete innovation

Complete innovation , This is a highly demanding innovation , There are generally two levels , One is interdisciplinary innovation , Development of a theory .

We can see from the situation in recent years , commonly , Masters with higher requirements , It will involve interdisciplinary algorithm innovation , such as , We apply some physical formulas to biochemical formulas , Combined with innovation , Or use a set of mathematical formulas , To study some liberal arts problems .

Theoretical development , This is the most demanding innovation , That is, there is no ready-made theoretical basis , Need us to study , Summarize relevant theories , General students' projects will not involve this requirement , It is usually some complex engineering projects or doctoral projects , Based on some known conditions , Derive some new formulas , To represent a process or phenomenon . This is very difficult .

Technology
Daily Recommendation