Ashraf O. Nassef

A hybrid GA/heuristic approach to the simultaneous scheduling of machines and automated guided vehicles

In this paper, the problem of simultaneous scheduling of machines and identical automated guided vehicles (AGVs) in flexible manufacturing systems is addressed with the objective of minimizing the makespan. This problem is composed of two interrelated decision problems: the scheduling of machines, and the scheduling of AGVs. Both problems are known to be NP-complete, resulting in a more complicated NP-complete problem when they are considered simultaneously. A new hybrid Genetic-algorithm/heuristic coding scheme is developed for the studied problem. The developed coding scheme is combined with a set of genetic algorithm (GA) operators selected from the literature of the applications of GAs to the scheduling problems. The algorithm is applied to a set of 82 test problems, which was constructed by other researchers, and the comparison of the results indicates the superior performance of the developed coding.

Design of prestressed concrete flat slab using modern heuristic optimization techniques

The main objective of the current study is to utilize the huge capabilities of modern heuristic search algorithms for structural design optimization of pre-stressed concrete slab; providing a general, flexible, and relatively easy to use tool for practicing engineers. A robust numerical tool integrating design, analysis, and optimization techniques is developed for this purpose. The tool utilizes the Finite Element method for the structural analysis of the system. Optimum values for the slab thickness, number and size of tendons, and tendon profile are found subject to design constraints imposed by the relevant code of practice. The objective function incorporates the cost of both concrete and prestressing tendons. Although the objective function is simple and monotonic, the optimization problem is quiet challenging due to the complexity and nonlinearity of the constraints in addition to the discreteness of some of the variables. As a demonstration problem, the optimum design of a square prestressed flat slab is sought. Direct search methods, heuristic optimization techniques such as Genetic Algorithms, and multi-objective optimization techniques are considered. Results indicated that search should be conducted along a constraint boundary. It is suggested to divide the design variables into two groups and to define a second objective function representing the distance away from the constraint. Optimization should be carried out for both objective functions (cost and distance) simultaneously. Using the suggested procedure, optimum design is found more efficiently.

Finite element model updating approach to damage identification in beams using particle swarm optimization

The use of vibration-based techniques in damage identification has recently received considerable attention in many engineering disciplines. While various damage indicators have been proposed in the literature, those relying only on changes in the natural frequencies are quite appealing since these quantities can conveniently be acquired. Nevertheless, the use of natural frequencies in damage identification is faced with many obstacles, including insensitivity and non-uniqueness issues. The aim of this article is to develop a viable damage identification scheme based only on changes in the natural frequencies and to attempt to overcome the challenges typically encountered. The proposed methodology relies on building a finite element model (FEM) of the structure under investigation. An improved particle swarm optimization algorithm is proposed to facilitate updating the FEM in accordance with experimentally determined natural frequencies in order to predict the damage location and extent. The method is tested on beam structures and was shown to be an effective tool for damage identification.

A particle swarm-based genetic algorithm for scheduling in an agile environment

In this paper, genetic algorithms are applied to the scheduling of a manufacturing system that is designed to support an assembly-driven differentiation strategy in the context of agile manufacturing. The system consists of a single flexible machine followed by multiple identical assembly stations. The objective of the scheduling problem is to minimize the makespan. A modified version of the genetic algorithm, inspired by the particle swarm optimization approach, is applied to the problem in addition to the general application of genetic algorithms. The objective is to investigate the potential that the particle swarm optimization concepts may have in improving the performance of genetic algorithms when applied to the chosen problem. The performance of these algorithms is compared to existing heuristics in the literature. A 2^3 factorial experiment, replicated twice, is used to compare the performance of the various approaches and identify the significant factors that affect the average percentage deviation from a lower bound. The results show that both versions of genetic algorithms applications outperform the existing heuristics in many instances and provide schedules that are shorter by as much as 15.5% in the cases considered. In addition, the modified application of genetic algorithms outperforms the regular application with shorter schedules by as much as 3.6% in many instances.

Surrogate Function of Post-Tensioning Cable Forces for Cable-Stayed Bridges

Inclined stay cables in cable-stayed bridges are post-tensioned in order to counteract the effect of the deck dead load. The evaluation of the post-tensioning forces is tackled in previous literature using either of the iterative unit load method, the force equilibrium method, or by using optimization search methods. All of the above mentioned methods are computationally intensive which hamper their further use in an optimal design algorithm of the bridge. This paper presents a simple surrogate function that can be used to evaluate post-tensioning cable forces in semi-cable stayed bridge under the action of the dead load. The post-tensioning forces for several configurations of the semi-fan cable-stayed bridges are evaluated using genetic algorithms, and are approximated using a simple polynomial function. The developed surrogate functions are used to study the effect of the main span length of the bridge as well as of the pylon height on the distribution and magnitude of the post-tensioning cable forces.

Stability of Y stiffeners in ship plating under uniaxial compressive loads

Abstract Ship plates are stiffened using different stiffeners. In this paper, a Y stiffener is considered and investigated. A Y stiffener–plate combination model is used to represent the stiffened panel. Our contribution includes characterizing the local instability (buckling) of Y stiffeners in stiffened panels under the action of uniaxial compressive loads. The mathematical derivations have been carried out to find the elastic buckling coefficient for the web of the T-part of the Y stiffener under suitable boundary conditions. Then, the critical value of the buckling stress has been calculated. Using the value of the critical stress and the assumption of uniform stress distribution, the buckling load is calculated for the Y stiffener–plate combination model. Using curve fitting of the analytically obtained results, approximate expressions for calculation of the elastic buckling coefficient of the T-part of the Y stiffener are obtained. These approximate expressions enable designers to calculate easily the elastic buckling coefficient from which the critical buckling stress of the T-part is obtained.

Database for the optimum design of semi-fan composite cable-stayed bridges based on genetic algorithms

The optimum design solution of cable-stayed bridges is a complicated task. The large number of design parameters, strict design constraints imposed by design codes, highly geometrically nonlinear behaviour and robust effect of post-tensioning cable forces make traditional design methods not capable of achieving this task. In this study, an optimisation technique that integrates a finite element model, a genetic algorithm and polynomial functions for evaluating post-tensioning cable forces is adopted to develop a database for the optimum design of cable-stayed bridges. The database describes the variations of the optimum design parameters, including the main span length, height of the pylon, number of stay cables and cross-sectional dimensions of all elements with the total length of the bridge. The study concerns the optimisation of three-span composite cable-stayed bridges with semi-fan cable arrangements. The study covers bridge lengths ranging from 250 to 700 m. The database is presented in the form of simple design charts and tables. Such database is useful in the preliminary design and cost estimation for this type of bridges.

Design Optimization of a Vehicle B-Pillar Subjected to Roof Crush Using Mixed Reactive Taboo Search

∗ corresponding author ABSTRACT The primary obstacle in automated design for crashworthiness is the heavy computational resources required during the optimization processes. Hence it is desirable to develop efficient optimization algorithms capable of finding good solutions without requiring too many model simulations. This paper presents an efficient mixed discrete and continuous optimization algorithm, Mixed Reactive Taboo Search (MRTS), and its application to the design of a vehicle B-Pillar subjected to roof crush conditions. The problem is sophisticated enough to explore the MRTS’ capability of identifying multiple local optima with a single optimization run, yet the associated finite element model (FEM) is not too large to make the computational resources required for global optimization prohibitive. The optimization results demonstrated that a single run of MRTS identified a set of better designs with smaller number of simulation runs, than multiple runs of Sequential Quadratic Programming (SQP) with several starting points.

Design Optimization of Reverse Osmosis Water Desalination Systems via Genetic Algorithms

This paper explores the application of genetic algorithms (GA) for optimal design of reverse osmosis (RO) water desalination systems. While RO desalination is among the most cost and energy efficient methods for water desalination, optimal design of such systems is rarely an easy task. In these systems, salty water is made to flow at high pressure through vessels that contain semi-permeable membrane modules. The membranes can allow water to flow through, but prohibit the passage of salt ions. When the pressure is sufficiently high, water molecules will flow through the membranes leaving the salt ions behind and are collected in a fresh water stream. Typical system design variables for RO systems include the number and layout of the vessels and membrane modules, as well as the operating pressure and flow rate. This paper explores models for single and two-stage RO pressure vessel configurations. The number and layout of the vessels and membrane modules are regarded as discrete variables, while the operating pressures and flow rate are regarded as continuous variables. GA is applied to optimize the models for minimum overall cost of unit produced fresh water. Case studies are considered for four different water salinity concentration levels. In each of the studies, three different types of crossover are explored in the GA. While all the studied crossover types yielded satisfactory results, the crossover types that attempt to exploit design variable continuity performed slightly better, even for the discrete variables of this problem.Copyright

Studies on the design of reverse osmosis water desalination systems for cost and energy efficiency

This paper explores optimal design of reverse osmosis (RO) systems for water desalination. In these systems, salty water flows at high pressure through vessels containing semi-permeable membrane modules. The membranes can allow water to flow through, but prohibit the passage of salt ions. When the pressure is sufficiently high, water molecules will flow through the membranes leaving the salt ions behind, and are collected in a fresh water stream. Typical system design variables include the number and layout of the vessels and membrane modules, as well as the operating pressure and flow rate. This paper presents models for single and two-stage pressure vessel configurations. The models are used to explore the various design scenarios in order to minimize the cost and energy required per unit volume of produced fresh water. Multi-objective genetic algorithm (GA) is used to generate the Pareto-optimal design scenarios for the systems. Case studies are considered for four different water salinity concentration levels. Results of the studies indicate that even though the energy required to drive the RO system is a major contributor to the cost of fresh water production, there exists a tradeoff between minimum energy and minimum cost. An additional parametric study on the unit cost of energy is performed in order to explore future trends. The parametric study demonstrates how an increase in the unit cost of energy may shift the minimum cost designs to shift to more energy-efficient design scenarios.Copyright