Determining how to allocate resources in order to prevent and prepare for disruptions is a challenging task for homeland security officials. Disruptions are uncertain events with uncertain consequences. Resources that could be used to prepare for unlikely disruptions may be better used for other priorities. This chapter presents an optimization model to help homeland security officials determine how to allocate resources to prevent and prepare for multiple disruptions and how to allocate resources to respond to and recover from a disruption. In the resource allocation model, prevention reduces the probability of a disruption, and preparation and response both reduce the consequences of a disruption. The model is applied to the US Gulf Coast region and considers a Deepwater Horizon‐type oil spill and a hurricane similar to Hurricane Katrina.

Controlling variation is an important aspect of quality improvement. Deming distinguishes between common cause variation and special cause variation and argues that both types of variation frequently result from people participating in the process. Confusing common cause and special cause variation can lead to incorrect decisions. This article analyzes the impact of an individual's ability to distinguish between common cause and special cause variation by simulating a manufacturing system with several human operators and a production manager. We use a recognition primed decision (RPD) model to simulate how human operators and the production manager would interpret the variation and make decisions to reduce the variation. A shared mental model with the RPD framework describes the interactions between different operators and the production manager. Results from this simulation demonstrate the importance of distinguishing between common cause and special cause variation, especially when problems occur at bottleneck points in the manufacturing system.

Engineering economics courses often require students to take time-constrained, in-class exams in which they solve problems by hand, possibly referring to interest rate tables. Many students rely on partial credit to successfully pass exams. Outside of the classroom, professionals rely on computers to solve engineering economics problems, which raises the question of whether engineering economics courses are correctly assessing student performance. This article describes the study of a large engineering economics class using a non-conventional testing method. Student performance was evaluated using online testing modules with a stringent passing criterion, and the tests could be taken multiple times. The questions for each testing attempt were pulled from a database so that students received a new question every time. We compare the performance of students who were assessed using traditional methods with the performance of students assessed with these online testing modules. Our analysis shows that, overall, students who were assessed using the online testing modules earned better grades than students who were assessed via traditional methods. The analysis also discusses several benefits and drawbacks to using online assessments compared with traditional methods. The online assessment method could be useful in large engineering courses that are formula-based.

Due to the increasing occurrence of disruptions across our global society, it has become critically important to understand the resilience of different socio-economic systems, i.e., to what extent those systems exhibit the ability both to resist a disruption and to recover from one once it occurs. In order to characterize this ability, however, one must be able to quantitatively measure the relative level of resilience that a given system displays in response to a disruptive event. Such a measurement should be easily understandable and straightforward to implement, but it should also utilize a consistent frame of reference so that one can properly compare the relative performance of different systems and assess the relative effectiveness of different resilience investments. With this in mind, this paper presents an improved approach for measuring system resilience that supports better decision making by providing both consistency and flexibility across different contexts. The theoretical basis for the approach is developed first, and then its advantages and limitations are illustrated in the context of several different practical examples.

The average cost of natural disasters and damage to the U.S. economy has increased each year from approximately $35 billion in 1980 to $300 billion in 2017. This increase in the cost of natural disasters could be due to an increase in the strength and frequency of natural disasters and/or growth in the U.S. economy. We forecast the cost of natural disasters by fitting probability distributions to the historical cost of billion-dollar disasters. We model the cost of natural disasters based on all weather-related natural disasters that cost more than $1 billion since 1980 and based only on those natural disasters that cost more than $1 billion that occurred in the past 20 years. Using the data from 1980 to 2018, the model forecasts the annual expected cost to be $52 billion. However, if only the recent disaster data is used to the fit the model, we forecast the expected annual cost to be $93 billion.

Porcine reproductive and respiratory syndrome (PRRS) is a disease caused by the PRRS virus (PRRSV) that has spread globally in the last 30 years and causes huge economic losses every year. This research aims to 1) investigate the relationship between the PRRSV detection in two age categories (wean-to-market and adult/sow farm), and 2) examine the extent to which the wean-to-market PRRSV positive rate forecasts the adult/sow farm PRRSV positive rate. The data we used are the PRRSV RNA detection results between 2007 and 2019 integrated by the US Swine Disease Reporting System project that represent 95% of all porcine submissions tested in the US National Animal Health Network. We first use statistical tools to investigate to what extent the increase in PRRSV positive submissions in the wean-to-market is related to the PRRSV increase in adult/sow farms. The statistical analysis confirms that an increase in the PRRSV positive rate of wean-to-market precedes the increase in the adult/sow farms to a large extent. Then we create the dynamic exponentially weighted moving average control charts to identify out-of-control points (i.e., signals) in the PRRSV rates for both wean-to-market and adult/sow farms. This control-chart-based analysis finds that 78% of PRRSV signals in the wean-to-market are followed by a PRRSV rate signal in the adult/sow farms within eight weeks. We expect that our findings will help the producers and veterinarians to justify and reinforce the implementation of bio-security and bio-contaminant practices to curb disease spread across farms.

In this paper, we investigate how to value an option to build a power plant when electricity demand fluctuates over time. Towards this aim, we first construct a transmission network, and obtain locational marginal prices for the network buses utilizing optimal power flow. Next, we construct a lattice model under the assumption that the demand fluctuation over time is represented by a geometric Brownian motion. Based on this demand lattice, we derive the economic consequences of costs to a bus with and without a power plant in a risk neutral world. These in turn will lead to the computation of the value of an option to build a power plant. This value of the option will be useful for the electric power planning as the bus with a higher value of this option indicates that the community in this bus is demonstrating a higher degree of potential need for such a power plant.

Designers should try to design systems that are resilient to adverse conditions during a system’s lifetime. The resilience of a system under time-dependent adverse conditions can be assessed by modeling the degradation and recovery of the system’s components. Decision makers in a firm should attempt to find the optimal design to make the system resilient to the various adverse conditions. A risk-neutral firm maximizes the expected profit gained from fielding the system, but a risk-averse firm may sacrifice some profit in order to avoid failure from these adverse conditions. The uniqueness of this paper lies in its model of a design firm’s risk aversion with a utility function or Value-at-Risk (VAR) and its use of that model to identify the optimal resilient design for the risk-averse firm. These risk-averse decision-making methods are applied to a design firm determining the resilience of a new engineered system. This paper significantly advances the engineering design discipline by modeling the firm’s appetite for risk within the context of designing a system that can fail due to degradation in the presence of adverse events and can respond to and recover from failure. Since the optimization model requires a complex Monte Carlo simulation to evaluate the objective function, we use a ranking and selection method and Bayesian optimization to find the optimal design. This paper incorporates the design of the wind turbine and the reliability and restoration of the turbine’s components for both risk-neutral and risk-averse decision makers. The results show that in order to make the system more resilient, risk-averse firms should pay a larger design cost to prevent catastrophic costs of failure. In this case, the system is less likely to fail due to the high resilience of its physical components.