Berna Dengiz

Local search genetic algorithm for optimal design of reliable networks

This paper presents a genetic algorithm (GA) with specialized encoding, initialization, and local search operators to optimize the design of communication network topologies. This NP-hard problem is often highly constrained so that random initialization and standard genetic operators usually generate infeasible networks. Another complication is that the fitness function involves calculating the all-terminal reliability of the network, which is a computationally expensive calculation. Therefore, it is imperative that the search balances the need to thoroughly explore the boundary between feasible and infeasible networks, along with calculating fitness on only the most promising candidate networks. The algorithm results are compared to optimum results found by branch and bound and also to GA results without local search operators on a suite of 79 test problems. This strategy of employing bounds, simple heuristic checks, and problem-specific repair and local search operators can be used on other highly constrained combinatorial applications where numerous fitness calculations are prohibitive.

Efficient optimization of all-terminal reliable networks, using an evolutionary approach

The use of computer communication networks has been rapidly increasing in order to: (1) share expensive hardware and software resources, and (2) provide access to main system from distant locations. The reliability and cost of these systems are important and are largely determined by network topology. Network topology consists of nodes and the links between nodes. The selection of optimal network topology is an NP-hard combinatorial problem so that the classical enumeration-based methods grow exponentially with network size. In this study, a heuristic search algorithm inspired by evolutionary methods is presented to solve the all-terminal network design problem when considering cost and reliability. The genetic algorithm heuristic is considerably enhanced over conventional implementations to improve effectiveness and efficiency. This general optimization approach is computationally efficient and highly effective on a large suite of test problems with search spaces up to 2/spl middot/10/sup 90/.

Multi-objective optimization of a stochastic assembly line balancing: A hybrid simulated annealing algorithm

This paper deals with multi-objective optimization of a single-model stochastic assembly line balancing problem with parallel stations. The objectives are as follows: (1) minimization of the smoothness index and (2) minimization of the design cost. To obtain Pareto-optimal solutions for the problem, we propose a new solution algorithm, based on simulated annealing (SA), called m_SAA. m_SAA implements a multinomial probability mass function approach, tabu list, repair algorithms and a diversification strategy. The effectiveness of m_SAA is investigated comparing its results with those obtained by another SA (using a weight-sum approach) on a suite of 24 test problems. Computational results show that m_SAA with a multinomial probability mass function approach is more effective than SA with weight-sum approach in terms of the quality of Pareto-optimal solutions. Moreover, we investigate the effects of properties (i.e., the tabu list, repair algorithms and diversification strategy) on the performance of m_SAA.

Optimization of buffer sizes in assembly systems using intelligent techniques

When the systems under investigation are complex, the analytical solutions to these systems become impossible. Because of the complex stochastic characteristics of the systems, simulation can be used as an analysis tool to predict the performance of an existing system or a design tool to test new systems under varying circumstances. However, simulation is extremely time consuming for most problems of practical interest. As a result, it is impractical to perform any parametric study of system performance, especially for systems with a large parameter space. One approach to overcome this limitation is to develop a simpler model to explain the relationship between the inputs and outputs of the system. Simulation metamodels are increasingly being used in conjunction with the original simulation, to improve the analysis and understanding of decision-making processes. In this study, an artificial neural network (ANN) metamodel is developed for the simulation model of an asynchronous assembly system and an ANN metamodel together with simulated annealing (SA) is used to optimize the buffer sizes in the system.

A branch and cut algorithm for the location-routing problem with simultaneous pickup and delivery

This paper addresses a location-routing problem with simultaneous pickup and delivery (LRPSPD) which is a general case of the location-routing problem. The LRPSPD is defined as finding locations of the depots and designing vehicle routes in such a way that pickup and delivery demands of each customer must be performed with same vehicle and the overall cost is minimized. We propose an effective branch-and-cut algorithm for solving the LRPSPD. The proposed algorithm implements several valid inequalities adapted from the literature for the problem and a local search based on simulated annealing algorithm to obtain upper bounds. Computational results, for a large number of instances derived from the literature, show that some instances with up to 88 customers and 8 potential depots can be solved in a reasonable computation time.

Simulation optimization using tabu search

Investigation of the performance and operation of complex systems in manufacturing or other environments, analytical models of these systems become very complicated. Because of the complex stochastic characteristic of the systems, simulation is used as a tool to analyze them. The trust of such simulation analysis usually is to determine the optimum combination of factors that effect the considered system performance. The purpose of this study is to use a tabu search algorithm in conjunction with a simulation model of a JIT system to find the optimum number of kanbans.

A General Neural Network Model for Estimating Telecommunications Network Reliability

This paper puts forth a new encoding method for using neural network models to estimate the reliability of telecommunications networks with identical link reliabilities. Neural estimation is computationally speedy, and can be used during network design optimization by an iterative algorithm such as tabu search, or simulated annealing. Two significant drawbacks of previous approaches to using neural networks to model system reliability are the long vector length of the inputs required to represent the network link architecture, and the specificity of the neural network model to a certain system size. Our encoding method overcomes both of these drawbacks with a compact, general set of inputs that adequately describe the likely network reliability. We computationally demonstrate both the precision of the neural network estimate of reliability, and the ability of the neural network model to generalize to a variety of network sizes, including application to three actual large scale communications networks.

Computer simulation of a PCB production line: metamodeling approach

Abstract This paper presents the results of two simulation models to investigate the effects of push and pull systems on a printed circuit board (PCB) manufacturing process at an electronics company in Ankara, Turkey. The current manufacturing process is a push system where the large batches of work-in-process are routed from one operation to another along a production line with 13 stations. It is shown that with the proposed pull system, the daily productivity can be increased by 12%. Through metamodeling with a regression metamodel analysis, the batch size optimization problem is considered for the proposed pull system of this factory producing PCBs which are used as main parts for many electronic devices such as tachometers, cash registers and printers. The regression metamodel greatly reduced the cost, time, and amount of the effort spent in analyzing the simulation. The model was validated and shown to provide good approximations to simulation results.

Buffer allocation and performance modeling in asynchronous assembly system operations: An artificial neural network metamodeling approach

This article investigates metamodeling opportunities in buffer allocation and performance modeling in asynchronous assembly systems (AAS). Practical challenges to properly design these complex systems are emphasized. A critical review of various approaches in modeling and evaluation of assembly systems reported in the recently published literature, with a special emphasis on the buffer allocation problems, is given. Various applications of artificial intelligence techniques on manufacturing systems problems, particularly those related to artificial neural networks, are also reviewed. Advantages and the drawbacks of the metamodeling approach are discussed. In this context, a metamodeling application on AAS buffer design/performance modeling problems in an attempt to extend the application domain of metamodeling approach to manufacturing/assembly systems is presented. An artificial neural network (ANN) metamodel is developed for a simulation model of an AAS. The ANN and regression metamodels for each AAS are compared with respect to their deviations from the simulation results. The analysis shows that the ANN metamodels can successfully be used to model of AASs. Consequently, one concludes that practising engineers involved in assembly system design can potentially benefit from the advantages of the metamodeling approach.

Optimal Design of Reliable Computer Networks: A Comparison of Metaheuristics

In many computer communications network design problems, such as those faced by hospitals, universities, research centers, and water distribution systems, the topology is fixed because of geographical and physical constraints or the existence of an existing system. When the topology is known, a reasonable approach to design is to select components among discrete alternatives for links and nodes to maximize reliability subject to cost. This problem is NP-hard with the added complication of a very computationally intensive objective function. This paper compares the performance of three classic metaheuristic procedures for solving large and realistic versions of the problem: hillclimbing, simulated annealing and genetic algorithms. Three alterations that use local search to seed the search or improve solutions during each iteration are also compared. It is shown that employing local search during evolution of the genetic algorithm, a memetic algorithm, yields the best network designs and does so at a reasonable computational cost. Hillclimbing performs well as a quick search for good designs, but cannot identify the most superior designs even when computational effort is equal to the metaheuristics.