Fang Zhi-geng

A conceptual framework for reverse logistics

As a powerful tool to solve the problems of resource shortage and environmental pollution, reverse logistics has been applied in more and more enterprises. Unfortunately, the theoretical studies on this field lag behind its engineering application. In order to solve this problem, a three-layer wheel-shaped framework, which is composed of a conceptual layer, an enabling layer and a support layer, is put forward. In the first layer, the connotation, elements, classification, basic characteristics and research contents of reverse logistics are addressed in detail. The enabling techniques are analyzed in the second layer. The basic facilities are discussed in support layer.

Using fractional GM(1,1) model to predict maintenance cost of weapon system

In order to improve the prediction performance for maintenance cost of weapon system, grey system model with the fractional order accumulation (FGM(1,1)) is used to predict the maintenance cost of weapon system with small sample. FGM(1,1) yielded the lower MAPE in the in-sample model and the much lower MAPE in the out-of-sample model, which means that FGM(1,1) can predict the memory process. FGM(1,1) can predict the maintenance cost of other weapon system.

Study on a grey Verhulst self-memory model and application

General grey Verhulst model for sigmoid process boasts good prediction accuracy. But the prediction accuracy is not high enough for fluctuant sigmoid process. In order to improve prediction accuracy, the grey Verhulst self-memory model was established based on grey system and self-memory principle. The model can recollect the historical data and get over grey Verhulst models shortages. Examples show that the suggested model can reflect the fluctuant trend of sigmoid process with satisfactory simulation and prediction precision.

Study on multiple attribute decision-making problem based on standard interval grey number and entropy weight

Multiple attribute decision-making means choosing the optimal alternative decision or arranging the decisions considering multiple attribute. In the solving of the multiple attribute decision-making problems, how to decide the weights of the attribute is very important. Till now, the acknowledging of the weight coefficient has gained a lot, which is given in conference [1] with the details of how to decide the weights of the evaluating index with entropy. However, due to the complexity of the decisions and the ambiguity of the indexes in practice, when evaluating a problem, the indexes are always represented by interval numbers, which mean an index is not a certain number, but in an interval. There is not enough research in the decision of the interval weights. So this article will change interval grey number into grey number in standard interval[2], and decide the indicator weights with entropy, in order to offer information to decision makers.

Study on Grey Graphical Evaluation Review Technique Network and its application

In this paper Grey Graphical Evaluation Review Technique (GERT in brief) Network is put forward and its model is built up, for the method of GERT Network requires accurate parameters of action, and however, people usually cannot find out exact value of parameters because of limited information of real life. By Grey GERT Network, many actions are easier to be described and discussed. This paper defines several types of Grey GERT Network and searches for the approach to solution. Two examples are also given to show the practical meaning.

The construction and application of process capability rapid determination model with poor information based upon the safest principle

Traditional process capability inspection is established on the basis of large sample and actually at the beginning mass product is difficult, so in the absence of large sample, process capability making use of small sample information can also be quickly judged. In this paper, we will fully dig the small sample information, process capability judgment will be converted to an optimization constraint problem to determine the coefficient of benchmark as decision variables, mean, variance, range, process capability, deviation coefficient as the dynamic constraint variables to build models and eventually the most reasonable decision optimization benchmark table will be found. The model and the benchmark table are applied to the practice test to judge its rationality, and through the final experiment to prove the validity of the model and the simplicity of benchmark table.

Investment bubble formation mechanism of Centipede Game based on grey integration and forwards induction

The investment process of decision-making in the economical system is a multi-agent-stage dynamic process, and bubble is generated, until it burst in the game among various economic entities. The paper constructs a Centipede Game network model of investment equilibrium-adjusted based on grey integration and forwards induction, and regulars the method framework based on grey integration and forwards induction; analyzing the formation mechanism of the bubble and invoking the concept system of Centipede Game network model, the paper divides the bubbles to integral bubble, structural bubble and frictional bubble according to the characteristics of bubble generation process; and verifies the validity of the model by taking Nanjings real estate market as an example.

Maximum entropy grey game model between human and nature

Taking the fact that nature chooses the game strategy in accordance with the laws of nature in the real world into account, this paper established of a new game model between human and nature under the gray space. In this model, the object player makes strategy choices based on the maximum entropy criterion, and the subject player follows the strategy of the object player, making strategy choices on the principle of maximizing its return, which points out the shortage of classic game between human and nature, extending the game to the gray space and expanding the application of game theory and grey system theory.

Predict China's per capita GDP based on ending-point optimized discrete grey (1, 1) model

In this paper, in view of the disadvantages of traditional discrete GM (1, 1) model, we propose the ending-point optimized discrete grey (1, 1) model (EODGM(1,1)) to improve prediction accuracy and give the concrete calculation formula. The novel model assumes that the sequence start iteration from the optimized ending point. We can get optimum initial iteration point by using optimization algorithm proposed in this paper. Because the ending point stands for the latest information, so the EODGM(1,1) is accord with the new information priority principle. We use the data of Chinas per capita gross domestic product (GDP) from 2001 to 2009 as an example, and the results show that the simulation accuracy of the EODGM(1,1) is superior to the traditional discrete GM (1,1) and ending-point fixed discrete grey model (EDFGM(1,1)). Furthermore, the one step prediction accuracy and two step prediction accuracy has more obvious advantages. The results show that this novel model is effective and applicable.

A GERT-EFL network model for the parametric configuration of maximum entropy

In this paper, a GERT-EFL network model (Graphical Evaluation Review Technique-Extracorporeal Feedback Loop) is constructed by the parametric configuration of maximum entropy. An analytic algorithm about a GERT-EFL network model for the parametric configuration of maximum entropy is provided. The model is applied to the real estate market. The mutual influence relationship of the real estate market each other is analyzed, then the sections of the extracorporeal feedback loop is also analyzed. The dynamic input-output shows condition of participants of the real estate market each other. The value-added of the resources flowing process is obtained.