General steps of comprehensive evaluation
1. Select the appropriate evaluation index according to the evaluation purpose. These indicators are very representative and distinctive, and can often be measured. The selection of evaluation indicators is mainly based on professional knowledge, that is, according to the theory and practice of related majors, the influence of each evaluation indicator on the results is analyzed, and those indicators that are representative, certain, different and independent are selected to form an evaluation index system.
2. According to the evaluation purpose, determine the relative importance of evaluation indicators in evaluating something, or the weight of each indicator; 3. Reasonably determine the evaluation grade and boundary of each single index;
4. Select the appropriate comprehensive evaluation method according to the evaluation purpose and data characteristics, and establish a comprehensive evaluation model according to historical data;
5. Determine the grade and quantitative limit of multi-index comprehensive evaluation. In the application practice of comprehensive evaluation of similar things, the selected evaluation model is tested and constantly revised and supplemented to make it scientific, practical and advanced, and then popularized.
Question 2: What algorithms are needed to participate in mathematical modeling? 1. Monte Carlo method:
Also known as computer random simulation method, also known as statistical experiment method. You can test the correctness of your model by simulation.
2. Data processing, such as data fitting, parameter estimation and interpolation.
There are a lot of data to be processed in the competition, and the key to data processing lies in these methods, which are usually assisted by matlab and can also deal with many fitting problems in combination with graphics.
3. Programming problem algorithm:
Including linear programming, integer programming, multivariate programming, quadratic programming, etc. Many problems in the competition are related to planning. It can be said that many models can be reduced to a set of inequalities as constraints and several function expressions as objective functions. Solving such a problem is the key.
This kind of problem can generally be solved by lingo software.
4. Graph theory problems:
This paper mainly studies the algorithms of this kind of problems, including Dijkstra, Floyd, prime number, Bellman-Ford, maximum flow, binary matching and so on. It shouldn't be difficult for people familiar with ACM.
5. Problems in computer algorithm design;
The algorithm design includes: dynamic programming, backtracking search, divide and conquer, branch and bound method (solving integer solutions) and so on.
6. Three non-classical algorithms of optimization theory:
A) simulated annealing
B) neural network
genetic algorithm
7. Grid algorithm and exhaustive algorithm
8. Discretization method for continuous problems
Because computers can only deal with discrete problems, and most of the data in reality are continuous, it is necessary to discretize the continuous problems before solving them with computers.
For example, the idea of difference instead of differential and summation instead of integral are all common methods to discretize continuous problems.
9. Numerical analysis method
This paper mainly studies various numerical calculation methods for solving mathematical problems, especially the methods and algorithms suitable for computer implementation.
Including: numerical approximation of functions, numerical differentiation and numerical integration, numerical solutions of nonlinear regression, numerical algebra, numerical solutions of ordinary differential equations, etc.
Matlab is mainly used to solve this problem.
10. Image processing algorithm
This part mainly uses matlab for image processing.
Including showing pictures, explaining questions and so on.
Question 3: What are the common methods of mathematical modeling? Accumulate algorithms and models, do real questions, that's all.