Bayesian networks for information synthesis in complex systems

This thesis introduces a method of applying Bayesian Networks to combine information from a range of data sources for effective decision support systems. It develops a set of techniques in development, validation, visualisation, and application of Complex Systems models, with a working demonstration in an Australian airport environment. The methods presented here have provided a modelling approach that produces highly flexible, informative and applicable interpretations of a system's behaviour under uncertain conditions. These end-to-end techniques are applied to the development of model based dashboards to support operators and decision makers in the multi-stakeholder airport environment. They provide highly flexible and informative interpretations and confidence in these interpretations of a system's behaviour under uncertain conditions.

Incorporating complex train paths to an analysis of absolute capacity

Railways are an important mode of transportation. They are however large and complex and their construction, management and operation is time consuming and costly. Evidently planning the current and future activities is vital. Part of that planning process is an analysis of capacity. To determine what volume of traffic can be achieved over time, a variety of railway capacity analysis techniques have been created. A generic analytical approach that incorporates more complex train paths however has yet to be provided. This article provides such an approach. This article extends a mathematical model for determining the theoretical capacity of a railway network. The main contribution of this paper is the modelling of more complex train paths whereby each section can be visited many times in the course of a train’s journey. Three variant models are formulated and then demonstrated in a case study. This article’s numerical investigations have successively shown the applicability of the proposed models and how they may be used to gain insights into system performance.

Techniques to effectively buffer schedules in the face of uncertainties

Resource assignment and scheduling is a difficult task when job processing times are stochastic, and resources are to be used for both known and unknown demand. To operate effectively within such an environment, several novel strategies are investigated. The first focuses upon the creation of a robust schedule, and utilises the concept of strategically placed idle time (i.e. buffering). The second approach introduces the idea of maintaining a number of free resources at each time, and culminates in another form of strategically placed buffering. The attraction of these approaches is that they are easy to grasp conceptually, and mimic what practitioners already do in practice. Our extensive numerical testing has shown that these techniques ensure more prompt job processing, and reduced job cancellations and waiting time. They are effective in the considered setting and could easily be adapted for many real life problems, for instance those in health care. This article has more importantly demonstrated that integrating the two approaches is a better strategy and will provide an effective stochastic scheduling approach.

New graph-based algorithms to efficiently solve large scale open pit mining optimisation problems

In the mining optimisation literature, most researchers focused on two strategic-level and tactical-level open-pit mine optimisation problems, which are respectively termed ultimate pit limit (UPIT) or constrained pit limit (CPIT). However, many researchers indicate that the substantial numbers of variables and constraints in real-world instances (e.g., with 50-1000 thousand blocks) make the CPIT’s mixed integer programming (MIP) model intractable for use. Thus, it becomes a considerable challenge to solve the large scale CPIT instances without relying on exact MIP optimiser as well as the complicated MIP relaxation/decomposition methods. To take this challenge, two new graph-based algorithms based on network flow graph and conjunctive graph theory are developed by taking advantage of problem properties. The performance of our proposed algorithms is validated by testing recent large scale benchmark UPIT and CPIT instances’ datasets of MineLib in 2013. In comparison to best known results from MineLib, it is shown that the proposed algorithms outperform other CPIT solution approaches existing in the literature. The proposed graph-based algorithms leads to a more competent mine scheduling optimisation expert system because the third-party MIP optimiser is no longer indispensable and random neighbourhood search is not necessary.

Using genetic algorithm in profile-based assignment of applications to virtual machines for greener data centers

The increase in data center dependent services has made energy optimization of data centers one of the most exigent challenges in today's Information Age. The necessity of green and energy-efficient measures is very high for reducing carbon footprint and exorbitant energy costs. However, inefficient application management of data centers results in high energy consumption and low resource utilization efficiency. Unfortunately, in most cases, deploying an energy-efficient application management solution inevitably degrades the resource utilization efficiency of the data centers. To address this problem, a Penalty-based Genetic Algorithm (GA) is presented in this paper to solve a defined profile-based application assignment problem whilst maintaining a trade-off between the power consumption performance and resource utilization performance. Case studies show that the penalty-based GA is highly scalable and provides 16% to 32% better solutions than a greedy algorithm.

A queuing approach for inventory planning with batch ordering in multi-echelon supply chains

This paper presents stylized models for conducting performance analysis of the manufacturing supply chain network (SCN) in a stochastic setting for batch ordering. We use queueing models to capture the behavior of SCN. The analysis is clubbed with an inventory optimization model, which can be used for designing inventory policies . In the first case, we model one manufacturer with one warehouse, which supplies to various retailers. We determine the optimal inventory level at the warehouse that minimizes total expected cost of carrying inventory, back order cost associated with serving orders in the backlog queue, and ordering cost. In the second model we impose service level constraint in terms of fill rate (probability an order is filled from stock at warehouse), assuming that customers do not balk from the system. We present several numerical examples to illustrate the model and to illustrate its various features. In the third case, we extend the model to a three-echelon inventory model which explicitly considers the logistics process.

A mathematical framework for expanding a railway’s theoretical capacity

Analytical techniques for measuring and planning railway capacity expansion activities have been considered in this article. A preliminary mathematical framework involving track duplication and section sub divisions is proposed for this task. In railways these features have a great effect on network performance and for this reason they have been considered. Additional motivations have also arisen from the limitations of prior models that have not included them.

An integrated mathematical approach to measure the capacity of MMTS

This article focusses upon multi-modal transportation systems (MMTS) and the issues surrounding the determination of system capacity. For that purpose a multi-objective framework is advocated that integrates all the different modes and many different competing capacity objectives. This framework is analytical in nature and facilitates a variety of capacity querying and capacity expansion planning.

The credit risk(+) model with general sector correlations

We consider an enhancement of the credit risk+ model to incorporate correlations between sectors. We model the sector default rates as linear combinations of a common set of independent variables that represent macro-economic variables or risk factors. We also derive the formula for exact VaR contributions at the obligor level.

Bottleneck assignment problems under categorization

This paper considers two special cases of bottleneck grouped assignment problems when n jobs belong to m distinct categories (m < n). Solving these special problems through the available branch and bound algorithms will result in a heavy computational burden. Sequentially identifying nonopitmal variables, this paper provides more efficient methods for those cases. Propositions leading to the algorithms have been established. Numerical examples illustrate the respective algorithms.

Optimization models for expanding a railway’s theoretical capacity

Changing the topology of a railway network can greatly affect its capacity. Railway networks however can be altered in a multitude of different ways. As each way has significant immediate and long term financial ramifications, it is a difficult task to decide how and where to expand the network. In response some railway capacity expansion models (RCEM) have been developed to help capacity planning activities, and to remove physical bottlenecks in the current railway system. The exact purpose of these models is to decide given a fixed budget, where track duplications and track sub divisions should be made, in order to increase theoretical capacity most. These models are high level and strategic, and this is why increases to the theoretical capacity is concentrated upon. The optimization models have been applied to a case study to demonstrate their application and their worth. The case study evidently shows how automated approaches of this nature could be a formidable alternative to current manual planning techniques and simulation. If the exact effect of track duplications and sub-divisions can be sufficiently approximated, this approach will be very applicable.

Open-pit block sequencing optimization: A mathematical model and solution technique

This study presents a comprehensive mathematical formulation model for a short-term open-pit mine block sequencing problem, which considers nearly all relevant technical aspects in open-pit mining. The proposed model aims to obtain the optimum extraction sequences of the original-size (smallest) blocks over short time intervals and in the presence of real-life constraints, including precedence relationship, machine capacity, grade requirements, processing demands and stockpile management. A hybrid branch-and-bound and simulated annealing algorithm is developed to solve the problem. Computational experiments show that the proposed methodology is a promising way to provide quantitative recommendations for mine planning and scheduling engineers.