Optimization Algorithms Testing and Convergence by Using a Stacked Histogram

• Main Authors: ZAPLATILEK, K., TALPA, M., LEUCHTER, J.
• Format: Article
• Language: English
• Published: Stefan cel Mare University of Suceava 2011-02-01
• Series: Advances in Electrical and Computer Engineering
• Subjects: MATLAB; convergence; optimization; stacked histogram; stochastic system
• Online Access: http://dx.doi.org/10.4316/AECE.2011.01002
• ISSN: 1582-7445; 1844-7600

The article describes an original method of optimization algorithms testing and convergence. The method is based on so-called stacked histogram. Stacked histogram is a histogram with its features marked by a chosen colour scheme. Thus, the histogram maintains the information on the input digital sequence. This approach enables an easy identification of the hidden defects in the random process statistical distribution. The stacked histogram is used for the testing of the convergent quality of various optimization techniques. Its width, position and colour scheme provides enough information on the chosen algorithm optimization trajectory. Both the classic iteration techniques and the stochastic optimization algorithm with the adaptation were used as examples.