Research Of Improved Ant Colony Algorithm On Gasoline Blending Optimization

The gasoline blending process is very important for refineries, and the blending recipe determined the final profits. The blending recipe directly affects the gasoline pass rate, and the blending process is a typical nonlinear constrained function optimization problem. Thus, to get a satisfactory gasoline recipe is the difficulties link and key process for refinery researchers.It is difficult to obtain satisfying optimum solution by traditional methods. According to this problem, an improved ant colony algorithm, called Ant Colony Optimization algorithm with Crossover operator (ACOC), was presented. The proposed algorithm introduced crossover operator into the ant colony algorithm, and improved the global search ability. In the process of local searching, the ACOC applied the Hooke-Jeeves algorithm to improve the performance of optimization algorithms and improve the convergence speed. The simulation results show that the ACOC method is very effective and adaptive widely. The ideal gasoline recipe can be found quickly by the proposed method, and get the max profit with little quality index margin. What's more, the proposed method broads the application scope of ant colony algorithm.Finally, this paper introduces the near infrared analyzer (INR), and proposed the solution method of the online pipeline oil blending.