Scheduling problems are common in a wide range of real-world industries and strategies for tackling them can impact on profits significantly. The careful planning and precise timing of industrial events and processes can maximize the utilization of resources, improve efficiency and reduce costs. One important type of scheduling problem from the domain of maintenance planning is the Generator Maintenance Scheduling Problem (GMSP) in the power industry. This thesis makes three broad contributions. First, we introduce a set of 23 real-world instances of the problem of different characteristics and sizes, based on data collected from industry in Saudi Arabia. We show that the fitness landscapes of these instances are rugged and full of relatively poor local optima. Our initial experiments to optimize the instances using a simple evolutionary algorithm and some hill-climbers reveal the dominance of local search for this problem, and suggest that effort be concentrated on the development of more advanced local search algorithms. Secondly, we turn our attention to ensemble problem solving, another promising direction, and propose the use of selection methods (selectors) to evaluate and choose the constituent algorithms of algorithm portfolios. These selectors range in intricacy and the computational effort they require. We show that a selector based on "racing" methods from the metaheuristic tuning literature appears to offer the best trade-off between performance and cost of selection. Finally, we propose several operators for an Iterated Local Search (ILS) algorithm for GMSP taking close account of the problem constraints. To improve performance, we propose extensions to the basic ILS design. These include an ILS with restart strategy, an ILS with delta evaluation implementation, an ILS hybrid with a variable neighbourhood descent algorithm, and a portfolio of ILSs. Results show a superior and consistent performance of the portfolios in a smaller number of evaluations (especially when using communication between constituent components) compared to the performance of individual constituent algorithms.

Diesel, one of the main petroleum products, is widely used in industry and transportation. Only high quality diesel product can survive in the more and more competitive market. The optimization methodology for diesel production and management is critical to refineries' profitability. LP/MIP models have been applied in diesel blending planning and scheduling in the last decades. With the benefits of reducing the model scale and computing efforts, LP/MIP models lead to operation results with inaccurate property estimation and profit loss due to the accuracy loss in the linearisation of blending models. To improve model accuracy, more accurate property prediction models for diesel blending should be incorporated into the refinery planning and schedule methods to improve decision making procedure in the case of scheduling for diesel blending, where academic effort is almost absent. A model for planning of refinery diesel streams is developed to optimise the diesel production of a refinery. Nonlinear blending models are applied to calculate blending properties more precisely than conventional linear models. Due to the large number of equations and variables, it may be generated to an infeasible solution if the given initial points are not good enough. To avoid this situation, a solution algorithm is proposed. Based on the NLP planning model, a model for scheduling diesel blending is developed. In order to improve the model accuracy, nonlinear blending correlations are used, which lead to a complicated MINLP problem that cannot be solved by existing MINLP solver directly. A robust solution algorithm is proposed in this thesis to help optimizing the MINLP problem. A case study of diesel production blending scheduling is introduced to illustrate how to model a diesel blending scheduling problem and the efficient and reliability of the solution algorithm. Besides, the proposed MINLP model and the solution algorithm can be extensively applied to other processes in a refinery, such as gasoline blending. Once gasoline blending models are taken into account, the model can be modified to optimize the gasoline blending scheduling problem.

Unrelated parallel machines are machines that perform the same function but have different capacity or capability. Thus, the processing time of each job would be different on machines of different types. The scheduling environment considered is dynamic in both job release time and machine availability. Additionally, each job considered can have different weight, and due date. Split jobs are also considered in this research. The number of jobs that needs to be processed in split-modes is pre-determined and not part of the scheduling decision. Additional constraints are imposed on split jobs to ensure that the absolute difference in completion time of the split portions of a job is within a user-specified margin. These constraints are supported by the Just-In-Time manufacturing concept where inventory has to be maintained at a very low or zero level. The objective of this research is to minimize the sum of the weighted tardiness of all jobs released within the planning horizon. The research problem is modeled as a mixed (binary) integer-linear programming model and it belongs to the class of NP-hard problems. Thus, one cannot rely on using an implicit enumeration technique, such as the one based on branch-and-bound, to solve industry-size problems within a reasonable computation time. Therefore, a higher-level search heuristic, based on a concept known as tabu search, is developed to solve the problems. Four different methods based on simple and composite dispatching rules are used to generate the initial solution that is used by tabu-search as a starting point. Six different tabu-search based heuristics are developed by incorporating the different features of tabu search. The heuristics are tested on eight small problems and the quality of their solutions is compared to their optimal solutions, which are obtained by applying the branch-and-bound technique. The evaluation shows that the tabu-search based heuristics are capable of obtaining solutions of good quality within a much shorter time. The best performer among these heuristics recorded a percentage deviation of only 1.18%. The performance of the tabu-search based heuristics is compared by conducting a statistical experiment that is based on a split-plot design. Three sizes of problem structures, ranging from 9 jobs to 60 jobs and from 3 machines to 15 machines are used in the experiment. The results of the experiment reveal that in comparison to other initial-generation methods, the composite dispatching rule is capable of obtaining initial solutions that significantly accelerate the tabu search based heuristic to get to the final solution. The use of long-term memory function is proven to be advantageous in solving all problem structures. The long-term memory based on maximum-frequency strategy is recommended for solving the small problem structure, while the minimum-frequency strategy is preferred for solving medium and large problem structures. With respect to the use of tabu-list size as a parameter, the variable tabu-list size is preferred for solving the smaller problem structure, but the fixed tabu-list size is preferred as the size of the problems grows from small to medium and then large. / Graduation date: 2000

In this work, we first introduce a novel approach to the long term irrigation scheduling using Genetic Algorithms (GAs). We explore the effectiveness of GAs in the context of optimizing nonlinear crop models and describe application requirements and implementation of the technique. GAs were found to converge quickly to near-optimal solutions. Second, we analyze the relationship between GA control parameters (population size, crossover rate, and mutation rate) and performance. We identify a combination of population, mutation, and crossover which searched the fitness landscape efficiently. The results suggest that smaller populations are able to provide better performance at relatively low mutation rates. More stable outcomes were generated using low mutation rates. Without crossover the quality of solutions were generally impaired, and the search process was lengthened. Aside from crossover rate zero, no other crossover rates significantly differed. The behaviors observed for best, online, offline, and average performances were sensitive to the combined influences control parameters. Interaction among control parameters was strongly indicated. Finally, several adaptive penalty techniques are presented for handling constraints in GAs, and their effectiveness is demonstrated. The constant penalty function suffered from sensitivity to settings of penalty coefficients, and was not successful in satisfying constraints. The adaptive penalty functions utilizes violation distance based metrics and search time based scaling using generation or trials number, and fitness values to penalize infeasible solutions, as the distance from the feasible region or number of generations increases so does the penalty. They were quite successful in providing solutions with minimal effort. They adapt the penalty as the search continues, encouraging feasible solutions to emerge over the time. Adaptive approaches presented here are flexible, efficient, and robust to parameter settings. / Graduation date: 1997

Thesis (M.S.)--Rochester Institute of Technology, 2007. / Typescript. Includes bibliographical references (leaves 86-90).

Scheduling problems have been studied extensively in the literature but because they are so hard to solve, especially real-world problems, it is still interesting to find ways of solving them more efficiently. This thesis aims to efficiently solve a real-world scheduling problem by using a hybrid representation together with an optimisation algorithm. The aim of the hybrid representation is to allow the optimisation to focus on the parts of the scheduling problem where it can make the most improvement. The new approach used in this thesis to accomplish this goal, is the combination of simulation-based optimisation using genetic algorithms and dispatching rules. By using this approach, it is possible to investigate the effect of putting specified job sequences in certain machines and using dispatching rules in the other. The hypothesis is that the optimisation can use dispatching rules on non-bottleneck machines that have little impact on the overall performance of the line and some specified job sequences on bottleneck machines that are hard to be scheduled efficiently with dispatching rules. This would allow the optimisation to focus on the bottleneck machines and that would produce a more efficient search. The results from the case study shows it is a viable approach exceeding or equalling existing techniques. The hypothesis that the optimisation can focus its efforts is supported by a bottleneck analysis which corresponds with the experimental results from optimisations.

In Industrial automated process plants, the wired communication system is being replacing by the newly developed wireless communication system as it is easy to use, scalable, simple, reliable, and provide more flexibility for installing and operating process automation equipments. The WirelessHART network is becoming popular for wireless communication system in industrial automation plant system. However this network works on the TDMA bus arbitration technique. Each network device needs to be scheduled by the user (network operator) to allow the communications between the field devices and a gateway. Some companies like DUST have already manufactured the device hardware to implement this wireless communication system on industrial automation plant. A Simulator of this system is necessary to imitate the system performance on computer using the computer programming to simulate the result of each timeslot. Purpose of designing this simulator is to use the timeslots more efficiently and to offer collision free communications between network devices. It is may be time taking process to build the simulator in the beginning but at the end it provides cost and time effective solutions.  This report describes how the simulator has been designed for the system at various phases e.g. architecture, design, and implementation etc.

In order to achieve the high performance, we need to have an efficient scheduling of a parallelprogram onto the processors in multiprocessor systems that minimizes the entire executiontime. This problem of multiprocessor scheduling can be stated as finding a schedule for ageneral task graph to be executed on a multiprocessor system so that the schedule length can be minimize [10]. This scheduling problem is known to be NP- Hard.In multi processor task scheduling, we have a number of CPU’s on which a number of tasksare to be scheduled that the program’s execution time is minimized. According to [10], thetasks scheduling problem is a key factor for a parallel multiprocessor system to gain betterperformance. A task can be partitioned into a group of subtasks and represented as a DAG(Directed Acyclic Graph), so the problem can be stated as finding a schedule for a DAG to beexecuted in a parallel multiprocessor system so that the schedule can be minimized. Thishelps to reduce processing time and increase processor utilization. The aim of this thesis workis to check and compare the results obtained by Bee Colony algorithm with already generatedbest known results in multi processor task scheduling domain.

Sensorless control of the synchronous reluctance motor has been a topic of research for more than a decade, producing several successful methods to accomplish this goal. However, a technique that has been overlooked is the full-order nonlinear observer, which is essentially a software model of the motor driven by measurements from the actual motor. Presented in this thesis is the design, implementation, and experimental testing of a full-order observer-based sensorless control technique which requires only the phase current and voltage measurements that are typically available in standard three-phase inverters. A technique is also presented for calculating a table of observer feedback gains parameterized only by the steady-state motor speed. This allows a gain-scheduling observer to be implemented which, as shown using experiments, improves the transient response of the observer over a wide speed range. The sensorless controller consists of a full-order nonlinear observer coupled with an input-output linearization speed controller. The resulting controller was implemented in Simulink and executed on a dSPACE DS1103 real-time DSP board using the Real-Time Workshop extension to Simulink. A custom built three-phase IGBT inverter was used to interface the DSP to a 100 watt synchronous reluctance motor for laboratory testing. The resulting sensorless controller was able to successfully track a varying speed reference from 150 rpm to 1800 rpm with a tracking error under 5% for most of the speed range. At the lowest speeds, the tracking error begins to increase but the observer remains stable down to 150 rpm.

Recent work in air transit efficiency has increased en-route efficiency to a point that airport efficiency is the bottleneck. With the expected expansion of air transit it will become important to get the most out of airport capacity. Departure scheduling is an area where efficiency stands to be improved, but due to the complicated nature of the problem an optimal solution is not always forthcoming. A heuristic approach can be used to find a sub-optimal take-off order in a significantly faster time than the optimal solution can be found using known methods. The aim of this research is to explore such heuristics and catalog their solution characteristics. A greedy approach as well as a k-interchange approach were developed to find improved takeoff sequences. When possible, the optimal solution was found to benchmark the performance of the heuristics, in general the heuristic solutions were within 10-15% of the optimal solution. The heuristic solutions showed improvements of up to 15% over the first-in first-out order with a running time around 4 ms.