Research On Multithreading Solution Of High Order Equation For Univariate

To solve high order equation for univariate is the key in the field of scientific research and engineering calculation.Many methods can be used,such as Newton's method and its improved methods.Their convergence is affected by the initial value,and the number of solutions is missing,etc.In view of the real roots and complex roots of high order equation for univariate,it is meaningful to study the effective method.Multithreading can meet the requirement for computing speed and accuracy,so it plays a crucial role in solving complex and large scale problems.It has a strong flexibility,can significantly improve the calculation speed,and make full use of existing hardware resources to achieve the maximum computational efficiency.The thesis mainly focuses on the method of multithreading to solve the high order equation for univariate.The improved algorithm's computation speed can be greatly improved by the concurrency of multithreading.The contents are three parts:(1)Multithreading related content,including differences between thread and process,thread mechanism and multithreading principle,etc.(2)It focuses on the traditional iterative method for solving real roots of the equation.To analyze its disadvantages,the method of combining multithreading and Golden Section method is used.The global convergence is better with the golden section,and combining with the good concurrency of multithreading,it can effectively reduce the times of iterations,get all the solutions,and significantly improve calculation speed.(3)The splitting factor method is often used to solve complex roots.Its convergence is uncertain,and combination of multithreading and genetic algorithm is used.Based on the characterstics of self-organization adaptive and the global optimization ability,it is effective to ensure the convergence and get the all complex roots quickly.