1: \begin{abstract}

2: The LP-Newton method solves the linear programming problem (LP)

3: by repeatedly projecting a current point onto a certain relevant polytope.

4: In this paper, we extend the algorithmic framework of the LP-Newton method 

5: to the second-order cone programming problem (SOCP)

6: via a linear semi-infinite programming (LSIP) reformulation of the given SOCP.

7: In the extension, we produce a sequence by projection onto polyhedral cones

8: constructed from LPs obtained by finitely relaxing the LSIP.

9: We show the global convergence property of the proposed algorithm under mild assumptions,

10: and investigate its efficiency through numerical experiments

11: comparing the proposed approach with the primal-dual interior-point method for the SOCP.

12: \keywords{Second-order cone program \and

13: 	Semi-infinite program \and

14: 	Adaptive polyhedral approximation \and

15: 	LP-Newton method}

16: % \PACS{PACS code1 \and PACS code2 \and more}

17: % \subclass{MSC code1 \and MSC code2 \and more}

18: \end{abstract}