| Comprehensive evaluation is a systemic and complex work, which is one of the important means for people to know, understand, and influence things. It is not only a process of management cognition, but also a process of decision making, which has been widely used in many fields such as economy, society, science and technology, education, management and engineering practice. Comprehensive evaluation has important values in the theory and application, which is a significant method to understand the evaluated issue profoundly and objectively, a basis of decision making for ranking and selecting alternatives, a crucial support tool to improve control measures and optimize management process, and an important basis for executing management behavior like rewards and punishments. Furthermore, the management and decision making based on the comprehensive evaluation aims to serve the management practice, improve control measures, optimize management process and ultimately enhance management effects.Around the purpose and steps of evaluation, the construction of indicators system, the determination of weight and value, the collection and processing of data, the integration of kinds of evaluation information, the use of the evaluation results and some other aspects, some sophisticated theory and methodology system have been formed. However, during the course of synthesized evaluation research, some specific problems which worthy of further research still exist. The research of this paper mainly focused on the standardization of construction steps and the approaches of indicators system, combination weighting based on subjective and objective information, the dependence and feedback information structures under a symbiotic environment including both certain and uncertain information in a comprehensive evaluation problem, and the risk of nonuniformity and poor convergence in the first combination evaluation process and some specific problems. After deeply analyzing, the solution methods has been given in this paper, and the given methods has been validated through case studies. All of these work devote to enriching and developing comprehensive evaluation theory and its methodology, and enhancing the ability to solve the actual problems.The main research contents and innovative work in this paper are as followed:Firstly, a standardization process and the approach to construct indicators system are studied. The principle of "O-C-W-I-S-D" is presented, which means objective, complete, workable, independent, significant and dynamic. The five stages model is given, which includes preliminarily building, preliminarily screening, quantitatively screening, rationality verification and feedback testing. An approach to construct indicators system is studied based on ground theory, statistical method, maximizing deviation method and factor analysis method. The paper studies a case about how to build the indicators system of regional science & technology capability in China. The results of the study show that the approach not only meet the principle of "O-C-W-I-S-D", but also simplify the number of the indicators while reflecting the main original informationSecondly, a combination weighting method based on Analytic Hierarchy Process (AHP) and Maximizing Deviation Decision method is studied. This paper defines the concept of weight and distinguish the classification of it. The common subjective, objective and combination methods to determine the weights are respectively reviewed and analyzed. A weighting method based on AHP and Maximizing Deviation Decision method is put forward. Furthermore, the definition of the approach is described in detail, also the operational and solving process, Matlab codes are given. In order to validate the presented approach, the selection and ranking for multiple Unmanned Aerial Vehicles mission planning alternatives are discussed. The results show that the approach not only effectively integrate some subjective information such as knowledge, experience and personal preferences from experts and decision makers, and some objective information including certain data, but also effectively rank the alternatives and select the best one.Thirdly, the paper analyzes the dependence and feedback information structures under a symbiotic environment including both certain and uncertain information in a comprehensive evaluation problem. Based on the detailed description on the problems, the standard multi-levels distributed evaluation model faced on dependence and feedback information structures is established. Then, the solution process using Analytic Network Process (ANP) and Evidential Reasoning (ER) is given. The evaluation for China’s national innovation index from 2005 to 2012 is analyzed to examine the model and the solution. Depending on the supplement of some uncertain information, and the disposal of the dependence and feedback among different information, China’s national innovation index is more aligned with China’s economy trend, and reflect China’s science & technology innovation capacity from 2005 to 2012 more truly.Finally, in order to handle with the nonuniformity and poor convergence in the first combination evaluation process, this paper proposed the second combination evaluation theory, and futher researches its concept and methodology. Some common combination evaluation methods are summarized, such as mean method, fuzzy borda method, Copeland method and Maximizing Deviation Decision method. On this basis, this paper presents and defines the second combination evaluation, and further proposes the concept of the second combination evaluation and builds model about it. From the numerical examples, we can see that the result error of the second combination evaluation is far less than result from any single first combination evaluation. Therefore, we can draw the conclusion that the second combination evaluation has better combination results than each first combination evaluation.
|
PDF Full Text Request