Abstract:In this paper we consider the default risk models with multi-period stochastic covariates，incorporating the dynamics of macroeconomic，firm-specific and industry-specific covariates which were used to characterize the credit contagion between industries． We propose maximum likelihood estimators for parameters of default risk intensities，and then show the maximum likelihood estimator for term structure of conditional corporate default probabilities． Meanwhile，it is easy to construct the significant test for credit contagion effect in multi-period corporate default risk models by the maximum likelihood estimators． Finally，simulation studies are conducted to compare the performance of the working independence estimator to that of the weighted marginal likelihood estimator．

Abstract:Provided that the investors have different information processing abilities in segmented markets，we build a two-stage model which includes heterogeneous beliefs of investors，prove that heterogeneous beliefs and short sale restriction is one of the reasons which result in lower H-share price than A-share price，and empirically analyze the models using data from A-share market and H-share market． To some extent，our model theoretically explains the anomalies of China’s segmented capital markets and is helpful to recognize the microstructure and market characteristics of emerging China capital markets

Abstract:In recent years，there have been growing studies in financial area on the dynamics of short-term interest rates，since it is an indispensable tool for interest rate derivatives pricing and risk management． This paper extends the short-rate dynamics from two perspectives: Stochastic volatility and regime switching，respectively． A novel method called particle filter is selected to estimate parameters and unobservable state variables for our Markov switching stochastic volatility model． Then we take our empirical analysis on China’s shortterm inter-bank borrowing interest rate． The results show that the short-rate has both stochastic volatility behavior and regime switching characteristic，and the BS-MSSV model is best for describing the short-rate dynamics． Moreover，our study confirms that neglecting regime switching of volatility mean would lead to overestimating the volatility persistence and worsen the modeling performance．

Abstract:Unexpected events，which may disable the original distribution plan，occur occasionally in urban distribution systems． In order to solve the difficult problem，the author adopted the methodology of disruption management to construct a lexicographic multi-objective disruption management model to handle the disruption caused by an unexpected event in urban distribution systems． This model measured disruptions from the following three aspects: Customer dissatisfaction，operational cost and path deviation． Tabu search algorithm，which combines several kinds of neighborhood functions，was adopted to solve the model． The computational results of the model proved that，due to the tradeoff between all parties involved in the distribution system，the model is more effective than existing global rescheduling models． Moreover，it is more flexible and feasible than traditional models in the aspects of objective ranking and customer dissatisfaction measuring in practical applications．

Abstract:Under the assumption that demand function is convex and has the non-decreasing price elasticity，this paper analytically studies the losses of both social welfare and operational efficiency in decentralized supply chains with price-sensitive demand，respectively． We consider the following system: A single supplier sells homogenous products to multiple identical retailers at the same wholesale price such that the market selling price can be determined under quantity competition． By virtue of the concept PoA ( Price of Anarchy) ，the whole profit or social welfare in the decentralized supply chain are compared with that in the integrated supply chain or the social optimal model，respectively． Then，the upper bounds for the losses，independent of the specific forms of demand functions，are derived analytically in detail． It is found that compared with the case of stochastic demand with fixed selling prices and inventory decision consideration，the inefficiency in the decentralized supply chain with price-sensitive demand has obviously higher upper bound when inventory decision is not integrated． Furthermore，the upper bounds increase with respect to the largest gross promargin per product and decrease as the competition at the downstream become keen．

Abstract:It is a hot topic to use the online algorithm and competitive analysis to study the online leasing problem． Based on the study of online leasing of general equipment，the online leasing of depreciable equipment is discussed． Since randomized algorithms can boost up performance，the randomized strategies for online leasing of depreciable equipment with oblivious adversary is proposed both with and without consideration of interest rate，respectively，and the maximum of its optimal competitive ratio is also obtained respectively． The conclusion shows that the introduction of depreciation factor made competitive performance improved． The introduction also makes the model more practical for large-scale investment in equipment problem，and provides investor better theoretical basis for decision making． In addition，consideration of interest rates makes the competitive ratio decrease a little，that is，investor will take more prudent investment strategy when interest rate is taken into consideration

Abstract:In view of the characteristics of semi-structured customer needs ( SCNs) with fuzziness and difficulty to be processed and converted to precise product functional requirements ( FRs) ，an expressing and transforming approach for SCNs is studied． The mathematical description approach of SCNs and the formal description approach of SCN transforming problem are presented，and then the architecture of the SCN transformation is constructed． This architecture is composed of four steps: fuzzy semantic analyzing，intelligent fuzzy reasoning，weighted optimizing，and defuzzification． Precise product functional requirements satisfying SCNs are outputted at last． Finally，the practicability of the approach is illustrated by an application case．

Abstract:Aiming at the premature convergence problem in discrete particle swarm optimization algorithm，a novel cellular particle swarm optimization algorithm is proposed，which is based on the principles of cellular automata and discrete particle swarm optimization algorithm． Cellular and its neighbor are introduced into the algorithm to maintain the swarm’s diversity and the algorithm uses evolutionary rule of cellular in local optimization to avoid local optima． Simulated tests of multi-dimensional knapsack problem and comparisons with other algorithms show the algorithm is feasible and effective and the algorithm has strong global optimization ability