Search

Dimensionality reduction, the process of reduce the number of dimension of the original feature set, plays an important role in a wide variety of contexts such as classification, Prediction and clustering. It is common to introduce the dimensionality reduction prior to the subsequent data mining tasks in the classical static data. However, this aforehand option can not capture the essence of datum because of the inherent time variety and one-pass constraint of data stream. In this case, dimensionality reduction should interact with the evolution of stream data with time elapsed. We introduced the interaction between dimensionality reduction in the time decaying high dimensional stream environment and propose the on-line fractal dimensionality reduction technique. Our performance experiments over a number of real and synthetic data sets illustrate the effectiveness and efficiency provided by our approach.

Design of vertical section of rolling down zone on hump affects rolling condition of car, and efficiency, ability and safety of shunting operating. Taking three factors, say, minimizing rolling time difference between front car and the latter, height of hump and the number of retarders as objectives, considering engineering technology criterion that should be satisfied, a Multi-objective optimization model is established and NSGA-II multi-objective genetic algorithm is adopted to obtain Patero Set in this paper. The model and algorithm are shown by a cast and both of them are reasonable.

Design of vertical section of hump pushing zone affects disconnecting capacity of marshalling yard. There are 2 factors being chosen to be the objective function, one is height difference between reception yard and hump crest, the other is the distance between the position of disconnecting the coupler of cars. Multi-objective optimization model is established. Meanwhile, Random weight multi-objective genetic algorithm is proposed. Pareto set method and reserving elite strategy is introduced as well. The model and algorithm are shown by a concrete example. As a result, both of them are reasonable.

The selection of path set has a great influence on stochastic traffic assignment. By considering path selection character of road users and traffic restriction in real road network, the concept of efficient paths is redefined. Moreover, a vertices-outdegree-based genetic algorithm (GA) is designed which adopts positive integer coding method. Furthermore, the probabilities of crossover and mutation are adjusted by adaptive algorithm and the process of selection is adopted simulated annealing algorithm to maintain the diversity and convergence of population. Thereby, this algorithm does not need to repair chromosome and can cover the shortages of priority - based GA. While finding efficient path sets of multi-OD (Origination-Destination) at the same time, this method can improve the computation efficiency by decoding simultaneously. At last, a case study is given for verifying the efficiency.

Determining the appropriate locations for inspection stations in a traffic network is a special kind of flow interception problem. It has very important applications in the fields of traffic management and police defense management. It is of great significance to research this problem. Based on the analysis of the characteristics of traffic flow, and on the assumption that the interarrival times of the cars follow Poisson distribution and the service time follows negative exponential distribution, this paper intends to research that with the case of the constraints of the expected car queue length, the expected car waiting time, the server busy rate and inspection station capacity taken into account, and based on queueing theory, a non-linear integer programming model for the problem of inspection station location is formulated. Finally, a numerical example is given.

With the development of ATIS(Advanced traveler information system), it is valuable to provide travel plans for the trip makers. For the bus trip makers, transfer times are always considered firstly. However, in most of cases, only the information about the bus routes, serving time, bus stops on these bus routes, etc, are offered ATIS. Considering the stochastic bus traveling time and transfer waiting time, the non-linear 0-1 programming model and the hyper- grahp model for optimal riding routes with transfer times limited under the random circumstance are formulated in this paper. And the improved priority-based encoding genetic algorithm is also designed. Finally, we give a numerical example.

Car-group-collection-and-delivery schedule of through trains distributed to several loading or unloading places is related directly to the car economic efficiency. Moreover large amount of calculation should be processed which is very difficult when the number of the loading or unloading places is large by use of the traditional algorithm. Taking the uncertainty of car group collection and delivery operation times and freight operating times into account, a fuzzy linear regression prediction model and an uncertain mathematical model of car-group-collection-and- delivery schedule for through trains are formulated and a Partheno-genetic algorithm (PGA) is put forward to resolve the model, besides, analyzed by using an example. The simulation results show that the algorithm is very efficient and has practical application.

In this paper, we present a genetic algorithm with a spanning tree representation for solving the bicriteria fixed charge transportation problem. First we consider the fixed charge transportation problem with a single objective function by the spanning tree-based genetic algorithm, and then extend this GA approach to solve the bicriteria problem. Due to the bicriteria program, the fitness function is constructed by dynamic scaling which normalizes the different data from the bicriteria region. The proposed genetic algorithm can find Pareto optimal solutions in the bicriteria space. Computational results will show the performance of the spanning tree-based genetic algorithm.

In this paper, we present an evolutionary program for solving the fuzzy MSTP (multicriteria solid transportation problem) in which the coefficients of objective function are represented as fuzzy numbers. The ranking fuzzy numbers with integral value are used in the evaluation and selection. The proposed algorithm is incorporated with problem-specific knowledge and conductive to find out the set of nondominated points in the criteria space based on decision maker's degree of optimism.

We present an evolutionary computation of the fuzzy mSTP (multicriteria solid transportation problem) in which the coefficients of the objective function are represented as fuzzy numbers. The ranking fuzzy numbers with integral value are used in the evaluation and selection. The proposed algorithm is incorporated with problem-specific knowledge and can find the set of nondominated points in the criteria space based on decision maker degree of optimism.

This paper describes an implementation of genetic algorithm to solve bicriteria solid transportation problem (BSTP), which is the extension of our previous work. The criteria space approach for bicriteria linear program is embedded in the structure of genetic algorithms (GA) for finding the nondominated points in the criteria space. The proposed algorithm is coded with C language. The computation results show that the proposed algorithms is efficient for solving the bicriteria solid transportation problem. >