Session ID: SA-G4-3 / Joint 4th International Conference on Soft Computing and Intelligent Systems and 9th International Symposium on advanced Intelligent Systems, September 17-21, 2008, Nagoya University, Nagoya, Japan

Health Sciences Centre (HSC) in Winnipeg is the major healthcare facility serving Manitoba, Northwestern Ontario, and Nunavut. An evaluation of HSC’s adult surgical patient flow revealed that one major barrier to smooth flow was their Operating Room (OR) scheduling system. This thesis presents a new two-stage elective OR scheduling system for HSC, which generates weekly OR schedules that reduce artificial variability in order to facilitate smooth patient flow. The first stage reduces day-to-day variability while the second stage reduces variability occurring within a day. The scheduling processes in both stages are mathematically modelled as multi-objective optimization problems. An attempt was made to solve both models using lexicographic goal programming. However, this proved to be an unacceptable method for the second stage, so a new multi-objective genetic algorithm, Nondominated Sorting Genetic Algorithm II – Operating Room (NSGAII-OR), was developed. Results indicate that if the system is implemented at HSC, their surgical patient flow will likely improve.

No description available.

Many areas in which computational optimisation may be applied are multi-objective optimisation problems; those where multiple objectives must be minimised (for minimisation problems) or maximised (for maximisation problems). Where (as is usually the case) these are competing objectives, the optimisation involves the discovery of a set of solutions the quality of which cannot be distinguished without further preference information regarding the objectives. A large body of literature exists documenting the study and application of evolutionary algorithms to multi-objective optimisation, with particular focus being given to evolutionary strategy techniques which demonstrate the ability to converge to desired solutions rapidly on many problems. Simulated annealing is a single-objective optimisation technique which is provably convergent, making it a tempting technique for extension to multi-objective optimisation. Previous proposals for extending simulated annealing to the multi-objective case have mostly taken the form of a traditional single-objective simulated annealer optimising a composite (often summed) function of the objectives. The first part of this thesis deals with introducing an alternate method for multiobjective simulated annealing, dealing with the dominance relation which operates without assigning preference information to the objectives. Non-generic improvements to this algorithm are presented, providing methods for generating more desirable suggestions for new solutions. This new method is shown to exhibit rapid convergence to the desired set, dependent upon the properties of the problem, with empirical results on a range of popular test problems with comparison to the popular NSGA-II genetic algorithm and a leading multi-objective simulated annealer from the literature. The new algorithm is applied to the commercial optimisation of CDMA mobile telecommunication networks and is shown to perform well upon this problem. The second section of this thesis contains an investigation into the effects upon convergence of a range of optimiser properties. New algorithms are proposed with the properties desired to investigate. The relationship between evolutionary strategies and the simulated annealing techniques is illustrated, and explanation of the differing performance of the previously proposed algorithms across a standard test suite is given. The properties of problems on which simulated annealer approaches are desirable are investigated and new problems proposed to best provide comparisons between different simulated annealing techniques.

Master of Engineering (Research) / In many multi-objective missions there are situations when actions based on maximum information gain may not be the `best' given the overall mission objectives. In addition to properties such as entropy, information also has value, which is situationally dependent. This thesis examines the concept of information value in a multi-objective mission from an information theory perspective. A derivation of information value is presented that considers both the context of information, via a fused world belief state, and a system mission. The derived information value is used as part of the objective function for control of autonomous platforms within a framework developed for human robot cooperative control. A simulated security operation in a structured environment is implemented to test both the framework, and information value based control. The simulation involves a system of heterogeneous, sensor equipped Unmanned Aerial Vehicles (UAVs), tasked with gathering information regarding ground vehicles. The UAVs support an e ort to protect a number of important buildings in the area of operation. Thus, the purpose of the information is to aid the security operation by ensuring that security forces can deploy e ciently to counter any threat. A number of di erent local controllers using information based control are implemented and compared to a task based control scheme. The relative performance of each is examined with respect to a number of performance metrics with conclusions drawn regarding the performance and exibility of information value based control.

The world of today is increasingly dependant on a functional, globalized economy. The defence and security establishments’ reliance on supplies and logistics is not new. First responders rely on many tools and systems that are critical to their endeavours. Somewhat disjoint at first glance, these domains share a common need for complex physical or logistical infrastructures such as power plants, ports, supply chains, to name a few examples.All of these are potentially vulnerable to attacks, disruptions, breakdowns, or other activities that disable the infrastructure and consequently cause important physical or economic damage. An obligation exists to protect these critical infrastructures and a decision support system that is able to detect, identify, and mitigate the risk of unwanted events would be invaluable in preventing the disastrous consequences of compromised infrastructure.This thesis explores the design and application of such a system. It starts with a pre-existing, actively researched risk management framework and proposes a methodology to apply it in new contexts, as well as contributions to provide the framework with the ability to solve new problems. Relevant case studies in critical infrastructure protection are presented, as well as applications of the developed methodology with the proposed modifications when suitable. Simulations, results, and insightful discussions are provided for each of the case studies. Finally, research trends, future work, and a conclusion are given, completing this thesis.

abstract: Neuron models that behave like their biological counterparts are essential for computational neuroscience.Reduced neuron models, which abstract away biological mechanisms in the interest of speed and interpretability, have received much attention due to their utility in large scale simulations of the brain, but little care has been taken to ensure that these models exhibit behaviors that closely resemble real neurons. In order to improve the verisimilitude of these reduced neuron models, I developed an optimizer that uses genetic algorithms to align model behaviors with those observed in experiments. I verified that this optimizer was able to recover model parameters given only observed physiological data; however, I also found that reduced models nonetheless had limited ability to reproduce all observed behaviors, and that this varied by cell type and desired behavior. These challenges can partly be surmounted by carefully designing the set of physiological features that guide the optimization. In summary, we found evidence that reduced neuron model optimization had the potential to produce reduced neuron models for only a limited range of neuron types. / Dissertation/Thesis / Doctoral Dissertation Neuroscience 2020

In the thesis of “Mohammed Aijaz Mohiuddin”, Engineering Nature-Inspired Heuristics for the Open Shortest Path First Weight Setting Problem, nature inspired heuristics were developed. Besides the existing two objectives, namely maximum utilization and the number of congested links, a third objective namely the number of unused links was used to formulate the fuzzy based objective function for the OSPFWS problem. The idea was to make use unused network links if any. Furthermore, a hybrid fuzzy based evolutionary Particle Swarm Optimization (FEPSO) algorithm was designed that harnessed evolutionary intelligence along with swarm intelligence. The proposed FEPSO algorithm was tested on different size test cases and its performance was mutually compared with other algorithms namely Simulated Annealing, Simulated Evolution, Particle Swarm Optimization, Weighted Aggregation Particle Swarm Optimization, Pareto-dominance Particle Swarm Optimization and Non-dominating Sorting Genetic Algorithm. Obtained results suggested the better performance of FEPSO among other algorithms over majority of test cases. / Thesis (PHD)--University of Pretoria, 2018. / Computer Science / PhD / Unrestricted

In this dissertation, we consider extensions of dynamic programming for combinatorial optimization. We introduce two exact multi-objective optimization algorithms: the multi-stage optimization algorithm that optimizes the problem relative to the ordered sequence of objectives (lexicographic optimization) and the bi-criteria optimization algorithm that simultaneously optimizes the problem relative to two objectives (Pareto optimization). We also introduce a counting algorithm to count optimal solution before and after every optimization stage of multi-stage optimization. We propose a fairly universal approach based on so-called circuits without repetitions in which each element is generated exactly one time. Such circuits represent the sets of elements under consideration (the sets of feasible solutions) and are used by counting, multi-stage, and bi-criteria optimization algorithms. For a given optimization problem, we should describe an appropriate circuit and cost functions. Then, we can use the designed algorithms for which we already have proofs of their correctness and ways to evaluate the required number of operations and the time. We construct conventional (which work directly with elements) circuits without repetitions for matrix chain multiplication, global sequence alignment, optimal paths in directed graphs, binary search trees, convex polygon triangulation, line breaking (text justification), one-dimensional clustering, optimal bitonic tour, and segmented least squares. For these problems, we evaluate the number of operations and the time required by the optimization and counting algorithms, and consider the results of computational experiments. If we cannot find a conventional circuit without repetitions for a problem, we can either create custom algorithms for optimization and counting from scratch or can transform a circuit with repetitions into a so-called syntactical circuit, which is a circuit without repetitions that works not with elements but with formulas representing these elements. We apply both approaches to the optimization of matchings in trees and apply the second approach to the 0/1 knapsack problem. We also briefly introduce our work in operation research with applications to health care. This work extends our interest in the optimization field from developing new methods included in this dissertation towards the practical application.

No description available.