The Improved TA And Markov Chain’s Application In The Enterprise Life Cycle

For combinatorial optimization problems, the simulated annealing algorithm can asymptotically converge to the true solution. The threshold accepting(TA) algorithm as the improved algorithm, accelerate the algorithm convergence speed. But when the complexity of problem increases, the threshold accepting algorithm still haven’t solve the contradiction between higher operation efficiency and accuracy of solutions. In order to improve the con-vergence of the algorithm、 reduce the computational cost and improve the accuracy of the solution, it is necessary to further study the threshold accepting algorithm.With the development of science technology and economics, the market competition is more and more fierce. The business strategy failure of Enterprise managers will lead to the company’s significant losses, even threat the company’s survival in the future and make the firm’s collapse, causing serious social losses. Therefore, familiar with the regularity of the de-velopment of enterprises, divide the stage of enterprise life cycle exactly, establish forecasting model to predict the development of the enterprise in advance, prepare for the business strategy, is an important issue to be solved for the enterprise management and sustainable development.Based on the existing literature, the work done in this paper is as follows:First of all, improving the threshold accepting algorithm for its limitations. Through a memory set in the improved threshold accepting algorithm to avoid the missing of the optimal solution when accept a bad solution; Set up a termination condition in the inner loop to improve the efficiency of the algorithm under getting the global optimal solution as far as possible. Before and after the threshold accepting algorithm for improving applied to the construction of the uniform design table. Analyze the time taken to get the design tables by the two algorithms, and make a significance test about the simulation results. The statistical results show that the improved threshold accepting algorithm can get a better solution and has a higher operation efficiency.Then, analyze the relationship between current state and previous state of the enterprise life cycle. Build a discrete time markov model for the state of the enterprise life cycle, predict the stage of the enterprise life cycle in the future. Determine whether the markov model meets positive recurrent、aperiodic、 irreducible, compute the steady state of the enterprise life cycle, compared with the actual situation of the development of Chinese enterprises to determine whether the result is reasonable and correct.At last, due to the different division method of the enterprise life cycle will get different results, this paper studies three main division methods:cash flow pattern method、 financial comprehensive index method、the selling rate curve trend method, use the matrix distance as a criterion to compare the three methods, the result shows that cash flow pattern method is more consistent with the actual situation.