توسعه شبکه عصبی‌تصمیم مبتنی بر الگوریتم ژنتیک برای ارزیابی ارجحیات‌ در مسائل تصمیم‌گیری چندهدفه

نوع مقاله : مقاله پژوهشی

نویسندگان

1 دانشجوی دکتری، گروه مهندسی صنایع، دانشگاه پیام نور، تهران، ایران

2 استادیار گروه مهندسی صنایع، بخش فنی و مهندسی، دانشگاه پیام نور، تهران، ایران

3 استاد دانشکده مهندسی صنایع، پردیس دانشکده های فنی، دانشگاه تهران، ایران

چکیده

بکارگیری شبکه‌های عصبی در تخمین و توصیف ساختار ارجحیت‌های تصمیم‌گیرنده، در حل مسائل تصمیم‌گیری چندهدفه در سال‌های اخیر بسیار مورد توجه قرار گرفته ‌است. شبکه عصبی تصمیم رویکردی نوین برای تخمین تابع مطوبیت تصمیم‌گیرنده در مسایل چندهدفه است. توسعه و بهبود روش‌های آموزش این نوع از شبکه‌ها، یافتن راه حل مرجح در مسایل چندهدفه، به خصوص مسایل با ابعاد بزرگ را تسهیل می‌نماید. در این مقاله، به منظور غلبه بر مشکلات روش‌های آموزشی مبتنی بر گرادیان و با هدف افزایش کارآیی شبکه عصبی تصمیم روش آموزشی آن توسعه داده شده ‌است و از الگوریتم ژنتیک برای آموزش این شبکه عصبی استفاده می‌شود. برای تنظیم پارامترهای شبکه عصبی تابع هزینه بهبود یافته‌ای پیشنهادی می‌شود و بر اساس این تابع هزینه پارامترهای شبکه عصبی بهینه‌سازی می‌شوند. رویکرد پیشنهادی در حل چندین مثال کاربردی بکارگرفته شده‌است که نتایج نشان می‌دهند که رویکرد پیشنهادی روشی کارآ به منظور تخمین تابع مطلوبیت –به‌خصوص غیرخطی- در حل مسائل تصمیم‌گیری چندهدفه می‌باشد. همچنین رویکرد پیشنهادی در تخمین توابع مطلوبیت مسائل چندهدفه گسسته نیز قابلیت بکارگیری دارد.

کلیدواژه‌ها


عنوان مقاله [English]

Development of a Genetic Algorithm-based Decision Neural Network for the Preference Assessment in Multi-objective Decision-Making Problems

نویسندگان [English]

  • Mohadeseh Nadershahi 1
  • Azam Dokht Safi Samghabadi 2
  • Reza Tavakkoli-Moghaddam 3
1 PhD. student, Department of Industrial engineering, Payame Noor University, Tehran, Iran
2 Department of Industrial Engineering, Payame Noor University (PNU), P. O. Box 19395-3697, Tehran, Iran
3 University of Tehran
چکیده [English]

The application of neural networks in estimating and describing the structure of decision makers' priorities has been very much considered in solving multi-objective decision problems in recent years. The neural network is a new approach to estimate the decision-making function of a multi-objective problem. Developing and improving the teaching methods of these types of networks facilitate to find the preferred solution in multi-dimensional issues, especially large-scale issues. In this paper, the educational method is developed to increase the efficiency of a neural network. In addition, a genetic algorithm is used to train this neural network. Furthermore, an improved cost function is proposed to adjust the parameters of the neural network and based on this function the cost parameters of the neural network are optimized. The efficiency of the proposed method is shown in solving several practical examples, including linear/nonlinear and discrete/continuous optimization problems. The efficiency of the proposed method is shown in solving several practical examples, including linear/nonlinear and discrete/continuous optimization problems.

کلیدواژه‌ها [English]

  • Multi-objective Decision Making
  • Utility Function
  • Neural Network Training
  • Genetic Algorithm
[1]      Yao, X., Evolving artificial neural network, Proc. IEEE, 1999, vol. 87, pp. 1423-1447.
[2]      Mitchell, M., An introduction to genetic algorithm (Complex Adaptive Systems), The MIT Press, 1998.
[3]      Wang, J., & Malakooti, B., A feed forward neural network for multiple criteria decision making, Computer & Operations Research, 1992, Vol. 19, No. 2, pp. 151-167.
[4]      Sun, M., Stam, A., Steuer, R. E., Interactive multiple objective programming problems using Tchebycheff programs and artificial neural networks, Computer & Operations Research, 2000, Vol. 27, No.7, PP.601-620.
[5]      Haykin, S., Neural networks: A comprehensive foundation. New York: Macmillan, 1994.
[6]      Hecht-Nielsen, R., Theory of the backpropagation neural networks, International Joint Conference on Neural Networks, 1989, pp. 593–611.
[7]      Chen, J., Lin, S., A neural network approach - decision neural network (DNN) for preference assessment, IEEE Transaction on Systems, Man and Cybernetics net, 2004, vol. 34, No. 2, pp.219-225.
[8]      Chen, J. & Lin, S., An interactive neural network-based approach for solving multiple criteria decision-making problems, Decision support systems, 2003, vol. 36, pp. 137-146.
[9]      Lang, K. J., Waibel, A.H., Hinton, G.E., A Time-delay neural network architecture for isolated word recognition, Neural Network, 1990, vol. 3, pp. 23-43.
[10]   Whitley, D., Starkweather, T., Bogart, C., Genetic algorithm and neural networks: optimizing connections and connectivity, Parallel Computing, 1990, vol. 14, pp. 347-361.
[11]   Stanley, K. O., Miikkulainen, R., Evolving neural network through augmenting topologies, Evolutionary computation, 2002, vol. 10, pp. 99-127.
[12]   Fels, S. S., Hinton, G. E., Glove-talk: A neural network interface between a data-glove and a speech synthesizer, Neural Networks, IEEE Transactions on Neural Network, 1993, vol. 4, pp.2-8.
[13]   Knerr, S., Personnaz, L., Dreyfus, G., Handwritten digit recognition by neural networks with single-layer training, Neural Networks, IEEE Transactions on Neural Network, 1992, vol. 3, pp.962-968.
[14]   Sutton, R. S., Two problems with backpropagation and other steepest-descent learning procedures for networks, in proceeding of the Eighth Annual Conference of the Cognitive Science Society, 1986, pp.823-831.
[15]   Whitley, D., The GENITOR algorithm and selection pressure: why rank-based allocation of reproductive trials is best, In: Proceedings of the 3th International Conference on Genetic Algorithms, 1989, PP. 116-123.
[16]   Porto, V. W., Fogel, D. B., Alternative neural network training methods, IEEE Expert, 1995, vol. 10, pp. 16-22.
[17]   Bartlett, P., Downs, T., Training a neural network with a genetic algorithm: Department of Electrical. University of Queensland, 1990.
[18]   Sexton, R., S., Dorsey, R., E., Johnson, J., D., Toward global optimization of neural networks: a comparison of the genetic algorithm and the backpropagation, Decision Support Systems, 1998, vol. 22, pp.171-185.
[19]   Topchy, A.P., lebedko, O. A., Neural network training by means of cooperative evolutionary search, Nuclear Instrument and Method in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 1997, vol. 389.
[20]   Kinnebrock, W., Accelerating the standard backpropagation method using a genetic approach, Neurocomputing, 1994, vol. 6, pp.583-588.
[21]   Chen, Y. M.,OConnell, R. M. Active power line conditioner with a neural network control, Industry Application, IEEE Transaction on, 1997, vol. 33, pp. 1131-1136.
[22]   Belew, R. K., McInerney, N. N. schraudolph, Evolving networks: using the genetic algorithm with connectionist learning, in Proc, Second Conference on Artificial Life, 1991, pp.511-547.
[23]   Hancock, P. J. B., Genetic Algorithms and permutation problems: a comparison of recombination operators for neural net structure specification, in Proceeding of the International Workshop on Combination of Genetic Algorithms and Neural Networks (COGANN-92), 1992,PP.108-122.
[24]   Montana, D. Davis, L., Training feed forward neural networks using genetic algorithms, In: Proceedings of the 11th International Joint Conference on Artificial Intelligence, 1989, pp. 762-767.
[25]   Deb, K., Anand, A., Joshi, D., A computationally efficient evolutionary algorithm for real-parameter optimization, Evolutionary Computation, 2002, vol. 10, pp. 371-395.
[26]   Cantu-paz, E., Kamath, C., An empirical comparison of combinations of evolutionary algorithms and neural network for classification problems, IEEE Transaction on Systems, Man and Cybernetics, Part B: Cybernetics, 2005, vol. 35, pp. 915-927.
[27]   Malakooti, B., Zhou, Y., Feed-forward artificial neural networks for solving discrete multiple criteria decision making problems, Management Science, 1994, Vol. 40, No. 11, pp. 1542–1560.
[28]    Sun, M., Stam, A., Steuer, R. E., Solving multiple objective programming problems using feed-forward artificial neural networks: The interactive FFANN procedure, Management Science, 1996, Vol. 42, No.6, pp. 835-849.
[29]   Malakooti, B., Zhou, Y., Approximating polynomial functions by feedforward artificial neural networks: Capacity, analysis, and design, Applied Mathematics and Computation, 1998, Vol. 90, No.1, pp. 27-51.
[30]   QU, Li Li, Chen, Yan, An interactive integrated MCDM based on FFAN and application in the selection of logistic center location, 14th International Conference on Management Science & Engineering, 2007.
[31]   Rumelhart, D. E., Geoffrey, E. H., Walliams, R.J., Learning representations by back-propagating errors, Nature, 1986, Vol. 323, pp.533-536,.
[32]   Mahajan, R., Kaur, G., Neural network using genetic algorithm, International Journal of computer applications, 2013,vol. 77, pp.6-11.