No abstract

Data envelopment analysis (DEA) is an important managerial tool for evaluating and improving the performance of decision making units. The existing DEA models are mostly limited to static environment using crisp data and are time-consuming and also have weak discriminating power. The aim of this work is to introduce a new fuzzy dynamic DEA model with missing values, which benefits from strengths of multi-objective modeling to overcome weakness and drawbacks of the classic DEA models. To check for quality and accuracy of the proposed model, this paper offers a comparative study to compare the discriminating power and computational efforts of the model with two problems in the literature taken as benchmarks. Also, this paper presents a real application of the fuzzy dynamic DEA model for assessing and ranking the level of performance for 56 railways around the globe using real data gathered from credible sources. The numerical case illustrates the model and the result may be used by railways to improve their performance efficiency compared to the best in the sample. Results for the comparative study and the real case reveal significant improvement in computational time and discriminating power.

A vehicular ad hoc network (VANET) is a network in which vehicles acting as dynamic nodes communicate with each other. A VANET is a suitable piece of infrastructure for developing intelligent transportation systems. Stable communication within a VANET leads to enhanced driver safety and better traffic management. The clustering technique, which organizes similar vehicles into similar groups, is a possible method for improving the stability of connectivity within a VANET. In this paper, two new clustering algorithms suited to the dynamic environment of a VANET are proposed. The multi-objective data envelopment analysis clustering algorithm as a mathematical clustering model and the ant system-based clustering algorithm as a meta-heuristic clustering model are introduced as algorithms for VANETs. A comparative simulation study in a highway environment is presented as well to evaluate the introduced methods and compare them with the most commonly used VANET clustering algorithms. The results show that the proposed algorithms offer improved stability and runtime along with relatively better performance than existing algorithms. Furthermore, the results show that in the VANET environment, the mathematical clustering model proposed herein yields better results than the meta-heuristic algorithm.

No abstract

No abstract