نوع مقاله : پژوهشی
نویسندگان
دانشکدهی فنی و مهندسی، گروه مهندسی صنایع، دانشگاه شاهد
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
Discussion on the quality of services and products has been an important issue during the last decades and usually is measured by ordinal scales. The ordinal logistic regression model is a method which can analyze and predict ordinal variables. This method can be applied to lots of cases in biology, engineering, criminology and etc. It is difficult to collect all experimentation data, so the optimum situations are not completely clear and hard to be known. Hence, optimization of the response of an ordinal logistic regression model is necessary to find the best set of controllable factors. The purpose of the optimization is to find the component of predictable factors with the best possible probability for the best rank of the response. The best possible probability must have the maximum probability considering other ranks calculated probabilities. We proposed heuristic and meta-heuristic algorithms for this purpose. The proposed meta-heuristic algorithm for optimizing ordinal logistic regression is entitled; OL-SA. This algorithm is based on simulated annealing, which is a well known meta-heuristic method. Simulated annealing is a popular local search, which is applied to solve discrete and continuous optimization problems. This method can escape from the local solution like other meta-heuristic methods. The proposed meta-heuristic method is implemented in the case study of a university. Student satisfaction of computer
site efficiency is evaluated, and the results are analyzed. Moreover, two numerical analyses have been performed for the proposed method. The first problem, with heuristic and meta-heuristic methods, is executed, and the
results are compared with the exact solution. It is recognized that the innovative meta-heuristic approach has a higher efficiency and accuracy. So, in light of the second problem, the proposed meta-heuristic is used for optimal solutions (optimal values of the control variables) and the predicted probability of the responses that have been put in rank. This meta-heuristic could be applied to some practical situations, such as predicting earthquake
resistance constructions at their verification stage by the designer.
کلیدواژهها [English]
)