Wang Ding-wei

An Ant Colony Optimization Algorithm for Multiple Travelling Salesman Problem

Multiple travelling salesman problem (MTSP) is a typical computationally complex combinatorial optimization problem, which is an extension of the famous travelling salesman problem (TSP). The MTSP can be generalized to a wide variety of routing and scheduling problems. It is known that classical optimization procedures are not adequate for this problem. The paper makes the attempt to show how the ant colony optimization (ACO) can be applied to the MTSP with ability constraint. In this paper, we compare it with MGA by testing several standard problems from TSPLIB. The computational results show that the proposed algorithm can find competitive solutions even not all of the best solutions within rational time, especially for large scale problems

Bidding strategies in sealed-bid reverse multi-attribute auctions

This study considers the reverse multi-attribute, or procurement auction in which the item for sale is defined by several quality attributes in addition to the price, the buyer is the auctioneer, and the sellers are the bidders. We focus on a variation of the first-price sealed-bid protocol termed first-score sealed-bid auction. We analyze a specific model for the protocol, and we provide optimal strategies for both the auctioneer and the bidders participating in the auctions. In addition, we conclude that the auctioneer can manipulate the auction to yield better outcome, however, as the number of the bidders participating in the auction increases, the buyer is motivated to announce a real utility function.

Model and algorithm for optimization of rescue center location of emergent catastrophe

The location of rescue centers is a key problem in optimal resource allocation and logistics in emergency response. We propose a mathematical model for rescue center location with the considerations of emergency occurrence probability, catastrophe diffusion function and rescue function. Because the catastrophe diffusion and rescue functions are both nonlinear and time-variable, it cannot be solved by common mathematical programming methods. We develop a heuristic embedded genetic algorithm for the special model solution. The computation based on a large number of examples with practical data has shown us satisfactory results.

Design and simulation on RFID-enabled hybrid Pull/Push control strategy for multi-echelon inventory of supply chain

RFID-enable electronic Kanban system can make the inventory status of the enterprises in supply chain remote real time visible. In consideration of the structural character and the function of the enterprises of the multi-echelon inventory system of supply chain distribution network, the RFID-Enable hybrid Push/Pull control strategy are designed. In this strategy, manufactures are controlled by the Push strategy, the distributions and retailers are controlled by the RFID-Enable Pull strategy based on RFID-enable electronic Kanban system. Three indicators that evaluate the performance of control strategies are presented, which include the total inventory cost, the total shortage loss and the inventory turnover rate. To test the performance of the RFID-Enable Push/Pull Hybrid control strategy, the hybrid strategy is compared with RFID-Enable Pull and Push control strategy. Base on the principle of discrete event system simulation, we design and realize the simulation model, and simulate many different structural supply chain systems controlled by the three kinds of control strategy mentioned above; the results of the simulation experiment demonstrate that the RFID-Enable Push/Pull Hybrid control strategy performs best among the three kinds of control strategy.

An approach to establish and evaluate pricing business process in electronic commerce

There are a lot of illogical factors in traditional pricing business process, such as low level of information, automation and poor net, which decrease pricing efficiency, especially in electronic commerce. To solve this problem, traditional pricing business process is analyzed, valued and reengineered with a method of business process quantitative analysis based on process algebra. Then a new pricing business process that is suitable for electronic commerce environment is established according to the principles of business process reengineering and the characteristics of pricing in electronic commerce environment. The evaluation criterions before and after process reengineering are investigated, and the results are satisfied. The results verify that the methods are practical.

Dynamic pricing model of digital knowledge commodity considering competition from congener commodity

The distinctive characteristics of digital knowledge commodity (DKC) make its pricing method different from traditional commoditypsilas. One hand, cost characteristic of DKC gives us more price choices. On the other hand, characteristics else of DKC bring us more difficulties to price. To solve this problem, the first-degree price discrimination is adopted, with respect to the stage in which the on-line sellerpsilas sale number does not exceed critical sale number. Then dynamic pricing model considering competition from congener commodity is built, in which commodity copyright cost, advertisement cost, competitorpsilas price, customerpsilas expectation of commodity and customer behavior are considered, for the sake of maximizing the sale and the profit of an on-line seller. In customer behavior model, the positive exterior influence of Web on customer purchase behavior is considered, and the influence of commodity brand, history sale number and price are quantized.

Application of Software Agent to e-Commerce Consumer Buying Support

Facing the flood of information and increasing competitive environment, the e-Commerce seller must make its software facility more customer-oriented, more intelligent, and build one-to-one marketing. A new e-Commerce model is presented to support consumer buying decision making. Some new intelligent technologies like software agent, neural computing, Analytical Hierarchy Process, data mining are used. Some application testing results prove the algorithms efficiency and the models practicality.

Modeling and Optimal Algorithm of Charge and Cast Scheduling of Steelmaking-Continuous Casting

On the basis of research on charge design and cast scheduling methods and process rules of steelmaking-continuous casting production scheduling in melting factory, we establish a constrained 0-1 integer quadratic programming model for optimal charge design and cast scheduling, and then propose the structure of PSO algorithm and the solving method to this model. And the simulation computation with practical data shows that both the model and the solving method proposed are effective.

Research of relationship between uncertainty and investment

This study focuses on revealing the relationship between uncertainty and investment probability through real option model involving investment critical trigger and project earning. Use of Matlab software on the experimental results showing that project earning volatility influences investment probability, led the authors to conclude that this notion is not always correct, as increasing uncertainty should have an inhibiting effect on investment, and that in certain situation, increasing uncertainty actually increases the investment probability and so, should have positive impact on investment.

Simulation on RFID-enable CONWIP control strategy for multi-echelon inventory of supply chain

Radio Frequency Identification (RFID) provides an efficient way to the innovation of inventory control strategies. In this paper, we design a RFID-Enable electronic Card system. And a RFID-Enable CONWIP control strategy is designed based on the system. Five criteria to evaluate the performance of control strategies are presented, which include the total inventory cost, the total shortage loss, the total operating cost, the satisfaction rate and the inventory turnover rate. To verify the performance of the RFID-Enable CONWIP control strategy, the strategy is compared with RFID-Enable (s, S) and Push control strategies. Three typical inventory systems of two kinds of products are simulated in the experiments. The results demonstrate the CONWIP strategy is the best one among the three kinds of strategies. It can reduce the total operating cost and enhance the service level for multi-echelon distribution networks in supply chains.