1: \begin{abstract}

2: \noindent In this paper, we generalize the classical extragradient algorithm for solving variational inequality problems

3: by utilizing non-null normal vectors of the feasible set. In particular, two conceptual algorithms are proposed and each of them has three different

4: variants which are related to modified extragradient algorithms.

5: 

6: \noindent Our analysis contains two main parts: The first part

7: contains two different linesearches, one on the boundary of the

8: feasible set and the other one along the feasible direction. The

9: linesearches allow us to find suitable halfspaces containing the

10: solution set of the problem. By using non-null normal vectors of the

11: feasible set, these linesearches can potentially accelerate the

12: convergence. If all normal vectors are chosen as zero, then some of

13: these variants reduce to several well-known projection methods

14: proposed in the literature for solving the variational inequality

15: problem. The second part consists of three special projection steps,

16: generating three sequences with different interesting features.

17: 

18: \noindent Convergence analysis of both conceptual algorithms is

19: established, assuming existence of solutions, continuity and a weaker

20: condition than pseudomonotonicity on the operator. Examples, on each

21: variant, show that the modifications proposed here perform better

22: than previous classical variants. These results suggest that our

23: scheme may significantly improve the extragradient variants.

24: \end{abstract}

25: