-
[show abstract]
[hide abstract]
ABSTRACT: Support vector regression (SVR) method is a novel type of learning machine algorithms, which is seldom applied to the development
of urban atmospheric quality models under multiple socio-economic factors. This study presents four SVR models by selecting
linear, radial basis, spline, and polynomial functions as kernels, respectively for the prediction of urban dust fall levels.
The inputs of the models are identified as industrial coal consumption, population density, traffic flow coefficient, and
shopping density coefficient. The training and testing results show that the SVR model with radial basis kernel performs better
than the other three both in the training and testing processes. In addition, a number of scenario analyses reveal that the
most suitable parameters (insensitive loss function ɛ, the parameter to reduce the influence of error C, and discrete level or average distribution of parameters σ) are 0.001, 0.5, and 2 000, respectively.
Key wordssupport vector regression-urban air quality-dust fall-socio-economic factors-radial basis function
Journal of Central South University of Technology 04/2012; 17(2):307-315. · 0.36 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: This study advances an integrated simulation, inference, and optimization method (ISIOM) for optimizing groundwater remediation systems. SIOM has the advantages of (i) automotive screening of potential explanatory variables (e.g., the pumping rates at various remediation wells), (ii) providing a flexible manner for investigating the linear, interactive, and quadratic effects of operating conditions on the benzene levels, and (iii) mitigating the computational efforts in optimization processes. The method is applied to a petroleum-contaminated site in western Canada for identifying the optimal remediation strategies under a given set of remediation durations and environmental standard levels. To examine the effect of pumping duration on contaminants removing efficiency, 4 duration options are considered including 5, 10, 15, and 20 years, respectively. The results indicate that the pumping duration would have effect on the optimized scheme. It is suggested that the 10-year duration would be more desirable than the 15-year one. The simulation results demonstrate that the peak benzene concentrations would be reduced to satisfy the environmental standard when the optimal remediation strategy is carried out.
Water Research 06/2008; 42(10-11):2629-39. · 4.86 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: The previous inexact mixed-integer linear programming (IMILP) method can only tackle problems with coefficients of the objective function and constraints being crisp intervals, while the existing inexact mixed-integer semi-infinite programming (IMISIP) method can only deal with single-objective programming problems as it merely allows the number of constraints to be infinite. This study proposes, an inexact mixed-integer bi-infinite programming (IMIBIP) method by incorporating the concept of functional intervals into the programming framework. Different from the existing methods, the IMIBIP can tackle the inexact programming problems that contain both infinite objectives and constraints. The developed method is applied to capacity planning of waste management systems under a variety of uncertainties. Four scenarios are considered for comparing the solutions of IMIBIP with those of IMILP. The results indicate that reasonable solutions can be generated by the IMIBIP method. Compared with IMILP, the system cost from IMIBIP would be relatively high since the fluctuating market factors are considered; however, the IMILP solutions are associated with a raised system reliability level and a reduced constraint violation risk level.
Waste Management 05/2008; 29(1):21-31. · 2.43 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: This study develops a nonlinear chance-constrained programming (NCCP) model for optimizing surfactant-enhanced aquifer remediation (SEAR) processes. The model can not only address the parameter uncertainty, but provide a reliability level for the identified optimal remediation strategy. To solve the NCCP model, stepwise cluster analysis (SCA) is used to create a set of proxy simulators for quantifying the relationships between operating conditions (i.e., pumping rate) and probabilities of benzene levels in violation of standard. Compared to conventional parametric inference techniques, SCA is independent of prior assumptions for model forms (e.g., linear or exponential ones) and capable of reflecting complex nonlinear relationships between operating conditions and probabilities. To alleviate the computational efforts in the optimization process, the generated proxy simulators are repeatedly called by simulated annealing (SA) to test the feasibility of each potential solution. The implicit of the optimal NCCP solutions is discussed through a laboratory-scale SEAR system where porosity and intrinsic permeability are treated as stochastic parameters. It is observed that well locations, environmental standards, reliability levels and remediation durations would have significant effects on optimal SEAR strategies. By comparing the predicted benzene concentration without and with remediation actions, it is indicated that the optimal SEAR process can guarantee the benzene concentration to meet the environmental standard with a high reliability level.
Environmental Science and Technology 04/2008; 42(6):2009-14. · 5.23 Impact Factor
