Construction Safety Risk Modeling and Simulation

By building on a genetic‐inspired attribute‐based conceptual framework for safety risk analysis, we propose a novel approach to define, model, and simulate univariate and bivariate construction safety risk at the situational level. Our fully data‐driven techniques provide construction practitioners and academicians with an easy and automated way of getting valuable empirical insights from attribute‐based data extracted from unstructured textual injury reports. By applying our methodology on a data set of 814 injury reports, we first show the frequency‐magnitude distribution of construction safety risk to be very similar to that of many natural phenomena such as precipitation or earthquakes. Motivated by this observation, and drawing on state‐of‐the‐art techniques in hydroclimatology and insurance, we then introduce univariate and bivariate nonparametric stochastic safety risk generators based on kernel density estimators and copulas. These generators enable the user to produce large numbers of synthetic safety risk values faithful to the original data, allowing safety‐related decision making under uncertainty to be grounded on extensive empirical evidence. One of the implications of our study is that like natural phenomena, construction safety may benefit from being studied quantitatively by leveraging empirical data rather than strictly being approached through a managerial perspective using subjective data, which is the current industry standard. Finally, a side but interesting finding is that in our data set, attributes related to high energy levels (e.g., machinery, hazardous substance) and to human error (e.g., improper security of tools) emerge as strong risk shapers.     

On tail behaviour of kth upper order statistics under fixed and random sample sizes via tail equivalence

For a fixed positive integer k, limit laws of linearly normalized kth upper order statistics are well known. In this article, a comprehensive study of tail behaviours of limit laws of normalized kth upper order statistics under fixed and random sample sizes is carried out using tail equivalence which leads to some interesting tail behaviours of the limit laws. These lead to definitive answers about their max domains of attraction. Stochastic ordering properties of the limit laws are also studied. The results obtained are not dependent on linear norming and apply to power norming as well and generalize some results already available in the literature. And the proofs given here are elementary.     

An integrated approach for water resources decision making under interactive and compound uncertainties

In recent years, the issue of water allocation among competing users has been of great concern for many countries due to increasing water demand from population growth and economic development. In water management systems, the inherent uncertainties and their potential interactions pose a significant challenge for water managers to identify optimal water-allocation schemes in a complex and uncertain environment. This paper thus proposes a methodology that incorporates optimization techniques and statistical experimental designs within a general framework to address the issues of uncertainty and risk as well as their correlations in a systematic manner. A water resources management problem is used to demonstrate the applicability of the proposed methodology. The results indicate that interval solutions can be generated for the objective function and decision variables, and a number of decision alternatives can be obtained under different policy scenarios. The solutions with different risk levels of constraint violation can help quantify the relationship between the economic objective and the system risk, which is meaningful for supporting risk management. The experimental data obtained from the Taguchi's orthogonal array design are useful for identifying the significant factors affecting the means of total net benefits. Then the findings from the mixed-level factorial experiment can help reveal the latent interactions between those significant factors at different levels and their effects on the modeling response.     

Extension of the pool-adjacent-violators algorithm

The pool-adjacent-violators algorithm (PAVA) is an efficient algorithm which converges in a finite number of steps. However, it has been applicable so far only in isotonic regression with the simple order. This report extends its applicability to other quadratic programming problems, including certain one-sided multivariate testing problems and concave regression problems.     

Selective Maintenance in System Reliability with Random Costs of Repairing and Replacing the Components

In this paper we consider the problem of determining the optimum number of repairable and replaceable components to maximize a system's reliability when both, the cost of repairing the components and the cost of replacement of components by new ones, are random. We formulate it as a problem of non-linear stochastic programming. The solution is obtained through Chance Constrained programming. We also consider the problem of finding the optimal maintenance cost for a given reliability requirement of the system. The solution is then obtained by using Modified E-model. A numerical example is solved for both the formulations.     

Stochastic correlation coefficient ensembles for variable selection

In this paper, we propose a novel Max-Relevance and Min-Common-Redundancy criterion for variable selection in linear models. Considering that the ensemble approach for variable selection has been proven to be quite effective in linear regression models, we construct a variable selection ensemble (VSE) by combining the presented stochastic correlation coefficient algorithm with a stochastic stepwise algorithm. We conduct extensive experimental comparison of our algorithm and other methods using two simulation studies and four real-life data sets. The results confirm that the proposed VSE leads to promising improvement on variable selection and regression accuracy.     

Mapping Invasive Species Risks with Stochastic Models: A Cross-Border United States-Canada Application for Sirex noctilio Fabricius

Nonindigenous species have caused significant impacts to North American forests despite past and present international phytosanitary efforts. Though broadly acknowledged, the risks of pest invasions are difficult to quantify as they involve interactions between many factors that operate across a range of spatial and temporal scales: the transmission of invading organisms via various pathways, their spread and establishment in new environments. Our study presents a stochastic simulation approach to quantify these risks and associated uncertainties through time in a unified fashion. We outline this approach with an example of a forest pest recently detected in North America, Sirex noctilio Fabricius. We simulate new potential entries of S. noctilio as a stochastic process, based on recent volumes of marine shipments of commodities from countries where S. noctilio is established, as well as the broad dynamics of foreign marine imports. The results are then linked with a spatial model that simulates the spread of S. noctilio within the geographical distribution of its hosts (pines) while incorporating existing knowledge about its behavior in North American landscapes. Through replications, this approach yields a spatial representation of S. noctilio risks and uncertainties in a single integrated product. The approach should also be appealing to decisionmakers, since it accounts for projected flows of commodities that may serve as conduits for pest entry. Our 30-year forecasts indicate high establishment probability in Ontario, Quebec, and the northeastern United States, but further southward expansion of S. noctilio is uncertain, ultimately depending on the impact of recent international treatment standards for wood packing materials.     

A hybrid genetic algorithm/mathematical programming approach to the multi-family flowshop scheduling problem with lot streaming

This paper presents a hybrid genetic algorithm/mathematical programming heuristic for the n-job, m-machine flowshop problems with lot streaming. The number of sublots for each job and the size of sublots are directly addressed by the heuristic and setups may be sequence-dependent. A new aspect of the problem, the interleaving of sublots from different jobs in the processing sequence, is developed and addressed. Computational results from 12 randomly generated test sets of 24 problems each are presented.     

Bayesian inference of asymmetric stochastic conditional duration models

This paper extends stochastic conditional duration (SCD) models for financial transaction data to allow for correlation between error processes and innovations of observed duration process and latent log duration process. Suitable algorithms of Markov Chain Monte Carlo (MCMC) are developed to fit the resulting SCD models under various distributional assumptions about the innovation of the measurement equation. Unlike the estimation methods commonly used to estimate the SCD models in the literature, we work with the original specification of the model, without subjecting the observation equation to a logarithmic transformation. Results of simulation studies suggest that our proposed models and corresponding estimation methodology perform quite well. We also apply an auxiliary particle filter technique to construct one-step-ahead in-sample and out-of-sample duration forecasts of the fitted models. Applications to the IBM transaction data allow comparison of our models and methods to those existing in the literature.     

Duality of the improved algebraic method (DIAM)

In this note we present a variant of the improved algebraic method (IAM) using a duality analysis to solve linear programming (LP) problems where more insights to the method are presented. When the coordinates of all vertices are computed, any feasible point can be expressed as a linear combination of the vertices. The objective function is expressed as a weighted sum of its evaluation at the feasible vertices and the optimal point is associated with the highest/lowest coefficient of the weighted sum.