Based Electromagnetism-like Mechanism Algorithm For Automatic Test Paper Problem

Electromagnetism algorithm is an intelligent algorithm using simulated physics the exclusive and attractive law among charged particles to solve real multi-dimensional problem. At present although it is applied to practical problems with electromagnetic algorithm is still small, and less than a decade since the electromagnetic method been proposed, the algorithm has a good performance in solving problems. In solving the process with electromagnetism algorithm, first initialize the population, and code individuals of population, to search outstanding individuals which have good performance in local, and then calculate the individual power, through the exclusive and attractive law between the charged individuals to form a new population, and then iteratively to find the optimal solution.The article gives some introductions for automatic test paper firstly, and compared a variety of advantages and disadvantages of the test paper algorithmic. Such as a little earlier the random test paper algorithmic, efficiency but low success rate of test paper relatively. Backtracking algorithm is an early test paper algorithm, it can accurately find the most optimal solution, but it can only be used in the item bank which is a small, because the algorithm traverse every solution, and when faced large item banks, its time is incalculable. Data mining and artificial intelligence give a number of solutions for automatic test paper of problem, but application lowly and practicality is not high on the whole. Genetic algorithms is used widely in automatic test paper, and showed good test paper, test paper process clear, and high success rate. Then it describe the original model of electromagnetism algorithmic, and improve local search with genetic factors, and a variety of functions were tested to prove that a effect of electromagnetic algorithm. Compare with genetic algorithm, the improved electromagnetic algorithm show an excellent performance in the multi-dimensional function.The automatic test paper system gives better results with improved algorithm. Before applying the algorithm, first introduce various parameters of test paper one by one, and define the corresponding expression and computational methods, which is an essential for test paper system. Good constraints to guide the algorithm in the right direction to complete the test paper. And then draw a flow chart of the test system of structure and function, describe The test paper processes and modules, then design test system of the database environment, were established and the corresponding relationship table structure, and finally the algorithm is applied to the test paper system and coding.