# American Institute of Mathematical Sciences

March  2020, 10(1): 13-21. doi: 10.3934/naco.2019029

## Numerical comparisons of smoothing functions for optimal correction of an infeasible system of absolute value equations

 Department of Applied Mathematics, Faculty of Mathematical Sciences, University of Guilan, Rasht, Iran

* Corresponding author: Saeed Ketabchi

Received  June 2018 Revised  April 2019 Published  May 2019

Optimal correction of an infeasible system of absolute value equations (AVEs), leads into a nonconvex and nonsmooth fractional problem. Using Dinkelbach's approach, this problem can be reformulated to form a single variable equation. In this paper, first, we have smoothed the equation by considering four important and famous smoothing functions (see[2,23]) and thus, to solve it, a smoothing-type algorithm based on the Difference of Convex (DC) algorithm-Newtown methods is proposed. Finally, the randomly generated AVEs were compared to find the best smoothing function.

Citation: Fakhrodin Hashemi, Saeed Ketabchi. Numerical comparisons of smoothing functions for optimal correction of an infeasible system of absolute value equations. Numerical Algebra, Control & Optimization, 2020, 10 (1) : 13-21. doi: 10.3934/naco.2019029
##### References:

show all references

##### References:
Performance profile of computing time of Algorithm 2 for the example 1
Performance profile of computing time of Algorithm 2 for the example 2
Numerical results for Example 1
 $\phi_1$ $\phi_2$ $\phi_3$ $\phi_4$ $n$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ 100 8.27e-17 5.52e-11 0.04 9.84e-17 1.20e-10 0.03 8.29e-17 1.16e-10 0.04 7.88e-17 1.95e-11 0.01 500 8.42e-17 1.21e-10 1.01 8.32e-17 3.04e-10 0.83 1.03e-16 3.52e-10 0.98 8.01e-17 1.35e-10 0.38 1000 8.25e-17 1.41e-10 5.37 9.34e-17 6.02e-10 4.30 1.05e-16 4.17e-10 4.83 8.78e-17 9.15e-11 2.13 1500 9.36e-17 2.01e-10 14.07 7.98e-17 7.67e-10 11.82 1.00e-16 3.98e-10 12.13 8.80e-17 2.24e-11 6.06 2000 8.99e-17 2.68e-10 30.14 9.17e-17 4.94e-10 23.72 9.95e-17 4.78e-10 26.56 9.30e-17 6.71e-11 11.48 2500 9.42e-17 3.10e-10 48.76 7.83e-17 5.78e-10 37.08 9.00e-17 7.10e-10 46.51 9.65e-17 1.18e-10 19.34 3000 8.84e-17 3.79e-10 85.25 9.87e-17 6.94e-10 66.18 8.06e-17 6.69e-10 78.62 9.62e-17 2.07e-10 35.11 3500 8.44e-17 4.18e-10 122.58 9.77e-17 8.11e-10 98.50 9.37e-17 7.81e-10 117.32 9.43e-17 2.82e-11 52.73 4000 8.89e-17 4.67e-10 162.58 9.82e-17 9.08e-10 134.06 9.35e-17 8.94e-10 158.81 8.23e-17 4.70e-11 73.29 4500 1.01e-16 5.23e-10 245.94 9.12e-17 9.87e-10 189.79 9.14e-17 1.02e-09 242.38 8.81e-17 6.53e-11 105.15 5000 9.17e-17 5.58e-10 289.93 8.92e-17 1.09e-09 241.79 9.31e-17 1.09e-09 288.33 9.07e-17 8.83e-11 122.53
 $\phi_1$ $\phi_2$ $\phi_3$ $\phi_4$ $n$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ 100 8.27e-17 5.52e-11 0.04 9.84e-17 1.20e-10 0.03 8.29e-17 1.16e-10 0.04 7.88e-17 1.95e-11 0.01 500 8.42e-17 1.21e-10 1.01 8.32e-17 3.04e-10 0.83 1.03e-16 3.52e-10 0.98 8.01e-17 1.35e-10 0.38 1000 8.25e-17 1.41e-10 5.37 9.34e-17 6.02e-10 4.30 1.05e-16 4.17e-10 4.83 8.78e-17 9.15e-11 2.13 1500 9.36e-17 2.01e-10 14.07 7.98e-17 7.67e-10 11.82 1.00e-16 3.98e-10 12.13 8.80e-17 2.24e-11 6.06 2000 8.99e-17 2.68e-10 30.14 9.17e-17 4.94e-10 23.72 9.95e-17 4.78e-10 26.56 9.30e-17 6.71e-11 11.48 2500 9.42e-17 3.10e-10 48.76 7.83e-17 5.78e-10 37.08 9.00e-17 7.10e-10 46.51 9.65e-17 1.18e-10 19.34 3000 8.84e-17 3.79e-10 85.25 9.87e-17 6.94e-10 66.18 8.06e-17 6.69e-10 78.62 9.62e-17 2.07e-10 35.11 3500 8.44e-17 4.18e-10 122.58 9.77e-17 8.11e-10 98.50 9.37e-17 7.81e-10 117.32 9.43e-17 2.82e-11 52.73 4000 8.89e-17 4.67e-10 162.58 9.82e-17 9.08e-10 134.06 9.35e-17 8.94e-10 158.81 8.23e-17 4.70e-11 73.29 4500 1.01e-16 5.23e-10 245.94 9.12e-17 9.87e-10 189.79 9.14e-17 1.02e-09 242.38 8.81e-17 6.53e-11 105.15 5000 9.17e-17 5.58e-10 289.93 8.92e-17 1.09e-09 241.79 9.31e-17 1.09e-09 288.33 9.07e-17 8.83e-11 122.53
Numerical results for Example 2
 $\phi_1$ $\phi_2$ $\phi_3$ $\phi_4$ $n$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ 100 8.85e-17 2.71e-11 0.03 1.04e-16 2.09e-11 0.03 9.03e-17 5.04e-11 0.03 9.36e-17 3.17e-11 0.02 500 8.30e-17 2.55e-11 1.06 8.38e-17 1.51e-11 1.01 8.70e-17 9.62e-11 1.05 8.91e-17 3.69e-10 0.48 1000 9.17e-17 1.63e-11 6.77 8.96e-17 5.38e-12 6.42 8.11e-17 6.93e-11 6.71 8.58e-17 2.82e-10 2.63 1500 8.66e-17 1.75e-11 18.49 8.76e-17 7.15e-12 18.40 8.94e-17 3.95e-11 18.49 8.76e-17 7.39e-11 7.69 2000 8.91e-17 1.52e-11 34.22 8.68e-17 6.17e-12 32.40 9.38e-17 4.12e-11 33.71 8.62e-17 2.07e-10 17.75 2500 9.45e-17 2.82e-11 63.48 9.18e-17 3.01e-12 60.78 7.82e-17 5.16e-11 62.92 8.66e-17 3.56e-10 33.42 3000 8.06e-17 1.70e-11 113.18 8.08e-17 8.29e-12 100.24 9.50e-17 3.59e-11 112.44 9.35e-17 6.29e-10 52.83 3500 8.56e-17 6.96e-12 169.34 1.01e-16 3.62e-12 162.04 1.02e-16 2.63e-11 162.79 7.56e-17 8.45e-11 81.62 4000 8.21e-17 1.34e-11 245.96 9.40e-17 3.91e-12 244.70 8.69e-17 3.41e-11 245.37 9.57e-17 1.38e-10 126.02 4500 7.84e-17 1.70e-11 351.13 9.31e-17 4.52e-12 313.81 9.19e-17 4.93e-11 328.67 8.42e-17 2.02e-10 175.69 5000 1.11e-16 3.03e-11 476.20 8.51e-17 4.20e-12 437.35 1.03e-16 5.77e-11 444.11 6.94e-17 2.67e-10 232.96
 $\phi_1$ $\phi_2$ $\phi_3$ $\phi_4$ $n$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ $Me$ $Merror$ $Mtime$ 100 8.85e-17 2.71e-11 0.03 1.04e-16 2.09e-11 0.03 9.03e-17 5.04e-11 0.03 9.36e-17 3.17e-11 0.02 500 8.30e-17 2.55e-11 1.06 8.38e-17 1.51e-11 1.01 8.70e-17 9.62e-11 1.05 8.91e-17 3.69e-10 0.48 1000 9.17e-17 1.63e-11 6.77 8.96e-17 5.38e-12 6.42 8.11e-17 6.93e-11 6.71 8.58e-17 2.82e-10 2.63 1500 8.66e-17 1.75e-11 18.49 8.76e-17 7.15e-12 18.40 8.94e-17 3.95e-11 18.49 8.76e-17 7.39e-11 7.69 2000 8.91e-17 1.52e-11 34.22 8.68e-17 6.17e-12 32.40 9.38e-17 4.12e-11 33.71 8.62e-17 2.07e-10 17.75 2500 9.45e-17 2.82e-11 63.48 9.18e-17 3.01e-12 60.78 7.82e-17 5.16e-11 62.92 8.66e-17 3.56e-10 33.42 3000 8.06e-17 1.70e-11 113.18 8.08e-17 8.29e-12 100.24 9.50e-17 3.59e-11 112.44 9.35e-17 6.29e-10 52.83 3500 8.56e-17 6.96e-12 169.34 1.01e-16 3.62e-12 162.04 1.02e-16 2.63e-11 162.79 7.56e-17 8.45e-11 81.62 4000 8.21e-17 1.34e-11 245.96 9.40e-17 3.91e-12 244.70 8.69e-17 3.41e-11 245.37 9.57e-17 1.38e-10 126.02 4500 7.84e-17 1.70e-11 351.13 9.31e-17 4.52e-12 313.81 9.19e-17 4.93e-11 328.67 8.42e-17 2.02e-10 175.69 5000 1.11e-16 3.03e-11 476.20 8.51e-17 4.20e-12 437.35 1.03e-16 5.77e-11 444.11 6.94e-17 2.67e-10 232.96
 [1] Xiaoqin Jiang, Ying Zhang. A smoothing-type algorithm for absolute value equations. Journal of Industrial & Management Optimization, 2013, 9 (4) : 789-798. doi: 10.3934/jimo.2013.9.789 [2] Carlo Sinestrari. Semiconcavity of the value function for exit time problems with nonsmooth target. Communications on Pure & Applied Analysis, 2004, 3 (4) : 757-774. doi: 10.3934/cpaa.2004.3.757 [3] Sophie Guillaume. Evolution equations governed by the subdifferential of a convex composite function in finite dimensional spaces. Discrete & Continuous Dynamical Systems, 1996, 2 (1) : 23-52. doi: 10.3934/dcds.1996.2.23 [4] Sanming Liu, Zhijie Wang, Chongyang Liu. Proximal iterative Gaussian smoothing algorithm for a class of nonsmooth convex minimization problems. Numerical Algebra, Control & Optimization, 2015, 5 (1) : 79-89. doi: 10.3934/naco.2015.5.79 [5] Qingsong Duan, Mengwei Xu, Yue Lu, Liwei Zhang. A smoothing augmented Lagrangian method for nonconvex, nonsmooth constrained programs and its applications to bilevel problems. Journal of Industrial & Management Optimization, 2019, 15 (3) : 1241-1261. doi: 10.3934/jimo.2018094 [6] Behrouz Kheirfam. A full Nesterov-Todd step infeasible interior-point algorithm for symmetric optimization based on a specific kernel function. Numerical Algebra, Control & Optimization, 2013, 3 (4) : 601-614. doi: 10.3934/naco.2013.3.601 [7] Liping Tang, Ying Gao. Some properties of nonconvex oriented distance function and applications to vector optimization problems. Journal of Industrial & Management Optimization, 2021, 17 (1) : 485-500. doi: 10.3934/jimo.2020117 [8] Ugo Bessi. The stochastic value function in metric measure spaces. Discrete & Continuous Dynamical Systems, 2017, 37 (4) : 1819-1839. doi: 10.3934/dcds.2017076 [9] Piermarco Cannarsa, Peter R. Wolenski. Semiconcavity of the value function for a class of differential inclusions. Discrete & Continuous Dynamical Systems, 2011, 29 (2) : 453-466. doi: 10.3934/dcds.2011.29.453 [10] Juan Li, Wenqiang Li. Controlled reflected mean-field backward stochastic differential equations coupled with value function and related PDEs. Mathematical Control & Related Fields, 2015, 5 (3) : 501-516. doi: 10.3934/mcrf.2015.5.501 [11] Yang Li, Liwei Zhang. A nonlinear Lagrangian method based on Log-Sigmoid function for nonconvex semidefinite programming. Journal of Industrial & Management Optimization, 2009, 5 (3) : 651-669. doi: 10.3934/jimo.2009.5.651 [12] Ariela Briani, Hasnaa Zidani. Characterization of the value function of final state constrained control problems with BV trajectories. Communications on Pure & Applied Analysis, 2011, 10 (6) : 1567-1587. doi: 10.3934/cpaa.2011.10.1567 [13] Nguyen Huy Chieu, Jen-Chih Yao. Subgradients of the optimal value function in a parametric discrete optimal control problem. Journal of Industrial & Management Optimization, 2010, 6 (2) : 401-410. doi: 10.3934/jimo.2010.6.401 [14] Na Zhao, Zheng-Hai Huang. A nonmonotone smoothing Newton algorithm for solving box constrained variational inequalities with a $P_0$ function. Journal of Industrial & Management Optimization, 2011, 7 (2) : 467-482. doi: 10.3934/jimo.2011.7.467 [15] Z.Y. Wu, H.W.J. Lee, F.S. Bai, L.S. Zhang. Quadratic smoothing approximation to $l_1$ exact penalty function in global optimization. Journal of Industrial & Management Optimization, 2005, 1 (4) : 533-547. doi: 10.3934/jimo.2005.1.533 [16] Zhiyou Wu, Fusheng Bai, Guoquan Li, Yongjian Yang. A new auxiliary function method for systems of nonlinear equations. Journal of Industrial & Management Optimization, 2015, 11 (2) : 345-364. doi: 10.3934/jimo.2015.11.345 [17] Lingfeng Li, Shousheng Luo, Xue-Cheng Tai, Jiang Yang. A new variational approach based on level-set function for convex hull problem with outliers. Inverse Problems & Imaging, 2021, 15 (2) : 315-338. doi: 10.3934/ipi.2020070 [18] Feng Ma, Jiansheng Shu, Yaxiong Li, Jian Wu. The dual step size of the alternating direction method can be larger than 1.618 when one function is strongly convex. Journal of Industrial & Management Optimization, 2021, 17 (3) : 1173-1185. doi: 10.3934/jimo.2020016 [19] Samira Shahsavari, Saeed Ketabchi. The proximal methods for solving absolute value equation. Numerical Algebra, Control & Optimization, 2021, 11 (3) : 449-460. doi: 10.3934/naco.2020037 [20] Jeremiah Birrell. A posteriori error bounds for two point boundary value problems: A green's function approach. Journal of Computational Dynamics, 2015, 2 (2) : 143-164. doi: 10.3934/jcd.2015001

Impact Factor: