Dominate gradient strategy based on pareto dominant and gradient method

In this paper, based on Pareto dominant theory and gradient method, we proposed a kind of dominate gradient strategy. Dominate gradient is an intervention rule of local search for solving algorithm design in order to solve practical problems. Through the analysis and discussion for two dimensional functions in detail, we defined the situations of real superior, real inferior, false inferior and false superior. By analyzing a function instance, we elaborated the situations for dominate gradient definition. In this paper, we extend the two dimensional functions to multi-dimensional functions, in order to put forward a new angle for seeking the gradient direction and optimizing the local search, and it is worth to be generalized and kept on further researches in solving practical problems.