We consider the problem of identifying the source location of a contaminant via analyzing changes in concentration levels observed by a sensor network in a river system. To address this problem, we propose a framework including two main steps: (i) pre-processing data; and (ii) training and testing a classification model. Specifically, we first obtain a data set presenting concentration levels of a contaminant from a simulation model, and extract numerical characteristics from the data set. Then, random forest models are generated and assessed to identify the source location of a contaminant. By using the numerical characteristics from the prior step as their inputs, the models provide outputs representing the possibility, i.e., a value between 0 and 1, of a spill event at each candidate location. The performance of the framework is tested on a part of the Altamaha river system in the state of Georgia, United States of America.

Multivariate CUSUM (MCUSUM) charts with fixed and variable scan radii have been used to detect increases of disease incidence counts in spatiotemporal biosurveillance. Biosurveillance through MCUSUM charts often requires intensive modeling of the monitored process, which can be challenging in cases involving a large number of monitored regions, an arbitrary marginal data distribution, and spatial correlation. Unlike other MCUSUM charts in the literature which assume a multivariate normal distribution for the disease count data, the MCUSUM chart we suggest in this paper is robust to non-normal distributions such as the Poisson. Our chart does not require extensive modeling of the underlying process and searches for its control limits via simple simulation and interpolation. While maintaining satisfactory accuracy of its control limits, the chart provides reliable performance under various data distributions, scan radii, and spatial correlation structures.

In this study, we consider a carbon emission cap-and-trade system in which the policymaker decides the cap for carbon emissions for each company and also has the power to regulate the carbon price in the carbon trading market for the purpose of minimizing total carbon emissions. We assume that there are n companies regulated in terms of carbon emissions by the policymaker, each of which emits carbon when producing its own product. After learning the carbon cap and carbon price regulated by the policymaker, each company makes simultaneous pricing and production decisions using the quick response strategy, and can trade some of its carbon emissions in the carbon market at the carbon price set by the policymaker, if the carbon emissions are below the cap. We model this non-cooperative game between the policymaker and companies as a Stackelberg game in which the policymaker is the leader and the companies are the followers. We show that there exists an equilibrium for the policymaker’s carbon pricing decisions and each company’s production and pricing decisions. From this equilibrium, we derive a carbon cap for the company at which the amount of traded carbon emissions is zero. This implies that some company’s production and pricing decisions, even under carbon emission restrictions, will be equal to those without the carbon emission restrictions. Also, we find that companies participating in the carbon cap-and-trade system would reduce their carbon emissions through reduced production, but can have a chance to improve profit through control of the product’s selling price.

This study addresses the problem of identifying the source location of a contaminant spill in a river system when a sensor network returns observations containing random measurement errors. To solve this problem, we suggest a new framework comprising three main steps: (i) spill detection, (ii) data preprocessing, and (iii) source identification. Specifically, we applied a statistical process control chart to detect a contaminant spill with measurement errors while keeping the false alarm rate at less than or equal to a user-specified value. After detecting a spill, we generated a nonlinear regression model to estimate a breakthrough curve of the observations and derive a characteristic vector of the estimated curve. Using the characteristic vector as an input, a random forest model was constructed with the sensor raising the first alarm. The model provides output values between 0 and 1 to represent the possibility of each candidate location being the true spill source. These possibility values allow users to identify strong candidate locations for the spill. The accuracy of our framework was tested on part of the Altamaha River system in Georgia, USA.