AUTHORS: Mohammed El Alaoui, Karim El Moutaouakil, Mohamed Ettaouil

Download as PDF

ABSTRACT: The Continuous Hopfield Networks (CHN) is a neural network tools which can be used to solve many problems like auto-memory and optimization problems. The dynamics of the CHN is described by differential equations system which is hard to solve analytically. That is why, the researchers use the Euler Cauchy method to calculate the CHN equilibrium point. Unfortunately, this method suffers from several problems, especially quality of the decision for a large step, sensibility to the slope function parameters and to the initial conditions. In this work, we use the well-known multi-step numerical method called Adams– Bashforth method, which is strong in terms of stability and performance, to calculate the equilibrium point of the CHN associated with the max stable problem. This method introduces an intermediary step to improve the Euler Cauchy method precision. The experimental results show that the (CHN+Adams-Bashforth) method produce a large max stable sets in comparison with the (CHN+Euler-Cauchy) method.

KEYWORDS: Continuous Hopfield Networks, Euler Cauchy method, Adams–Bashforth method, max-stable problem

REFERENCES:

[1] H.Asgari, Y .S. Kavian, A. Mahani, A systolic architecture for Hopfield neural networks, Procedia Technology, Vol.17, 2014, pp. 736– 741.

[2] G. Joya, M.A. Atencia, F. Sandoval, Hopfield neural networks for optimization: study of the different dynamics, Neurocomputing, 2002, pp. 219-237.

[3] Amanda C. Mathias, Paulo C. Rech, Hopfield neural network: The hyperbolic tangent and the piecewise-linear activation functions, Neural Networks, 2012, pp. 42-45.

[4] Tariq Samad, Paul Harper, High-order Hopfield and Tank optimization networks, Parallel Computing, 1990, pp. 287-292.

[5] Rachid Sammouda, Nuru Adgaba, Ameur Touir, Ahmed Al-Ghamdi, Agriculture satellite image segmentation using a modified artificial Hopfield neural network, Computers in Human Behavior, 2014, pp. 436-441.

[6] Rong Long Wang, Zheng Tang, Qi Ping Cao, A learning method in Hopfield neural network for combinatorial optimization problem, Neurocomputing, 2002, pp. 1021-1024.

[7] J.J. Hopfield, D.W. Tank, Neural computation of decisions in optimization problems, Biological Cybernetics, 1985, pp. 1-25.

[8] A.H. Gee and S.V.B. Aiyer, and R.W. Prager, An analytical framework for optimizing neural networks, Neural Networks, 1993, pp. 79-97.

[9] P.M. Talaván, J. Yàñez, The generalized quadratic knapsack problem, Neural Networks, 2006, pp. 416-428.

[10] J. Håstad, Clique is hard to approximate within 𝑛𝑛1−𝜀𝜀 , Acta Math,1999, pp. 105-142.

[11] C. Mannino and A. Sassano, An exact algorithm for the maximum cardinality stable set problem. Computational Optimization and Applications, 1994, pp. 243-258.

[12] K. Tatsumi, Y. Yagi, and T. Tanino, Improved projection Hopfield network for the quadratic assignment problem, Proceedings of the 41st SICE Annual Conference, 2002, pp.2295-2300.

[13] R. Carrahan and P. M. Pardalos, An exact algorithm for the maximum clique problem, Operations Research Letters, 1990, pp. 375- 382.

[14] G. Gruber and F. Rendl, Computational experience with stable set relaxations, SIAM J Opt, 2003, pp. 1014-1028.

[15] C. Friden, A. Hertz, and D. de Werra, Stabulus: a technique for finding stable sets in large graphs with tabusearch, Computing, 1989, pp.35-45.

[16] P.M. Talaván and J. Yàñez,A continuous Hopfield network equilibrium points algorithm, computers and operations research, 2005, pp.2179-2196.

[17] DIMACS. The Second DIMACS Implementation Challenge, ftp://dimacs.rutgers.edu/pub/challenge/graph/be nchmarks/ clique/.

[18] M. Ettaouil, M. Lazaar, K. El moutaouakil, K. Haddouch, A New Algorithm for Optimization of the Kohonen Network Architectures Using the Continuous Hopfield Networks, WSEAS TRANSACTIONS on COMPUTERS, 2013.

[19] J.C. Butcher, Numerical methods for ordinary differential equations in the 20th century, Journal of Computational and Applied Mathematics, 2000, pp. 1-29.

[20] Shichang Ma, Yufeng Xu, and Wei Yue, Numerical Solutions of a Variable-Order Fractional Financial System, Journal of Applied Mathematics, 2012.

[21] Shaban Gholamtabar, Nouredin Parandin, Numerical solutions of second-order differential equations by Adam Bashforth method, American Journal of Engineering Research, 2014, pp. 318-322.

[22] J.C. Chiou, S.D. Wu, On the generation of higher order numerical integration methods using lower order Adams-Bashforth and Adams-Moulton methods, Journal of Computational and Applied Mathematics, 1999, pp. 19-29.

[23] R.D. Brandt, Y. Wang, A.J. Laub, and S.K. Mitra, Alternative networks for solving the travelling salesman problem and the listmatching problem, proceedings of the International Conference on Neural Networks, 1988, pp. 333-340.

[24] Ferhat Kurtulmuş, İsmail Kavdir, Detecting corn tassels using computer vision and support vector machine, Expert Systems with Applications, 2014, pp. 7390–7397.

[25] Marc Demange, Tınaz Ekim, Bernard Ries, Cerasela Tanasescu, on some applications of the selective graph coloring problem, European Journal of Operational Research, 2015, pp. 307-314.

[26] A. Fischer, F. Fischer, G. Jäger, J. Keilwagen, P. Molitor, I. Grosse, Exact algorithms and heuristics for the Quadratic Traveling Salesman Problem with an application in bioinformatics, Discrete Applied Mathematics, 2014, pp. 97- 114.

[27] Yerim Chung, Marc Demange, The 0–1 inverse maximum stable set problem, Discrete Applied Mathematics, 2008, pp. 2501-2516.

[28] F.H. Clarkea, Yu.S. Ledyaevb, R.J. Stern, Asymptotic Stability and Smooth Lyapunov Functions, Journal of Differential Equations, 1998, pp. 69–114.