pedram Pourkarim guilani; Mani Sharifi; parham azimi; maghsoud Amiri
Abstract
Due to the high sensitivity in applying of electronic and mechanical equipment, creating any conditions to increase the reliability of a system is always one of the important issues for system designers. Hence, making academic models much closer to the real word applications is very attractive. In the ...
Read More
Due to the high sensitivity in applying of electronic and mechanical equipment, creating any conditions to increase the reliability of a system is always one of the important issues for system designers. Hence, making academic models much closer to the real word applications is very attractive. In the most studies in the reliability area, it is assumed that the failure rates of the system components are constant and have exponential distributions. This distribution and its attractive memory less property provide simple mathematical relationships in order to obtain the system reliability. But in real word problems, considering time-dependent failure rates is more realistic to model processes. It means that, the system components do not fail with a constant rate during the time horizon; but this failure rate changes over the time. One of the most useful statistical distributions in order to model the time-dependent failure rates is the Weibull distribution. This distribution is not a memory less one, so it was impossible to apply simple and explicit mathematical relationships as the same as exponential distributions for the reliability of a system. Therefore, researchers in this field have used simulation technique in these circumstances which is not an exact method to get near-optimum solutions. In this paper, for the first time, it is tried to obtain a mathematical equation to calculate the reliability function of a system with time-dependent components based on Weibull distribution. Also, in order to validate the proposed method, the results compared with exact solution that exists in literature.
shaghayegh Vaziri; Arash Zaretalab; Mani Sharifi
Abstract
Course timetabling problem is a weekly assignment a set of course and teacher to the time and space with considering a lot of hard and soft constraints in universities. In each semester, heads of educational institutes take too much time and effort to prepare a timetable by using trial and error method ...
Read More
Course timetabling problem is a weekly assignment a set of course and teacher to the time and space with considering a lot of hard and soft constraints in universities. In each semester, heads of educational institutes take too much time and effort to prepare a timetable by using trial and error method or last semester's timetable, although the rapid changing needs, resources and rules of each semester causes this method are not the perfect solutions. In this study, we design and develop a novel multi objective mathematical model which taking into account the preferences of students and teachers, Due to the complexity, we have benefited the metaheuristic algorithm to solve nonlinear model. Simulated Annealing algorithm is used to solve the mathematical model in two stages. In the first stage, the system automatically generates feasible solutions that will meet all the hard constraints. Then, the solutions are improved with spotting different neighborhood's structures. This collection is in the form of computer software application which is implemented the C# language programing and SQL database. This system is tested the data gathered by Azad University data and the results compared to the manual process showed the great progress is achieved. The entire system is flexible and easy to test different scenarios