Abdullah Konak

Multi-objective optimization using genetic algorithms: A tutorial

Multi-objective formulations are realistic models for many complex engineering optimization problems. In many real-life problems, objectives under consideration conflict with each other, and optimizing a particular solution with respect to a single objective can result in unacceptable results with respect to the other objectives. A reasonable solution to a multi-objective problem is to investigate a set of solutions, each of which satisfies the objectives at an acceptable level without being dominated by any other solution. In this paper, an overview and tutorial is presented describing genetic algorithms (GA) developed specifically for problems with multiple objectives. They differ primarily from traditional GA by using specialized fitness functions and introducing methods to promote solution diversity.

Estimation of all-terminal network reliability using an artificial neural network

The exact calculation of all-terminal network reliability is an NP-hard problem, with computational effort growing exponentially with the number of nodes and links in the network. During optimal network design, a huge number of candidate topologies are typically examined with each requiring a network reliability calculation. Because of the impracticality of calculating all-terminal network reliability for networks of moderate to large size, Monte Carlo simulation methods to estimate network reliability and upper and lower bounds to bound reliability have been used as alternatives. This paper puts forth another alternative to the estimation of all-terminal network reliability -- that of artificial neural network (ANN) predictive models. Neural networks are constructed, trained and validated using the network topologies, the link reliabilities, and a network reliability upperbound as inputs and the exact network reliability as the target. A hierarchical approach is used: a general neural network screens all network topologies for reliability followed by a specialized neural network for highly reliable network designs. Both networks with identical link reliability and networks with varying link reliability are studied. Results, using a grouped cross-validation approach, show that the ANN approach yields more precise estimates than the upperbound, especially in the worst cases. Using the reliability estimation methods of the ANN, the upperbound and backtracking, optimal network design by simulated annealing is considered. Results show that the ANN regularly produces superior network designs at a reasonable computational cost.

Redundancy allocation for series-parallel systems using a max-min approach

The redundancy allocation problem is formulated with the objective of maximizing the minimum subsystem reliability for a series-parallel system. This is a new problem formulation that offers several distinct benefits compared to traditional problem formulations. Since time-to-failure of the system is dictated by the minimum subsystem time-to-failure, a logical design strategy is to increase the minimum subsystem reliability as high as possible, given constraints on the system. For some system design problems, a preferred design objective may be to maximize the minimum subsystem reliability. Additionally, the max-min formulation can serve as a useful and efficient surrogate for optimization problems to maximize system reliability. This is accomplished by sequentially solving a series of max-min subproblems by fixing the minimum subsystem reliability to create a new problem. For this new formulation, it becomes possible to linearize the problem and use integer programming methods to determine system design configurations that allow mixing of functionally equivalent component types within a subsystem. This is the first time the mixing of component types has been addressed using integer programming. The methodology is demonstrated on three problems.

Multiple Weighted Objectives Heuristic for the Redundancy Allocation Problem

A new heuristic is proposed and tested for system reliability optimization. The multiple weighted objective heuristic is based on a transformation of the problem into a multiple objective optimization problem, and then ultimately, transformation into a different single objective problem. The multiple objectives are to simultaneously maximize the reliability of each individual subsystem. This is a logical approach because system reliability is the product of the subsystem reliabilities, so if they are maximized, the system reliability will also be high. This new formulation and associated heuristic are then based on solving a sequence of linear programming problems. It is one of the very few optimization approaches that allow for linear programming algorithms and software to be used for the redundancy allocation problem when mixing of functionally equivalent components is allowed. Thus, it represents an efficient solution method that relies on readily available optimization tools. The heuristic is tested on many example problems, and compared to competing solution approaches. Overall, the heuristic performance is observed to be very good on the tested problem, and superior to the max-min heuristic regarding both efficiency, and performance

A new mixed integer programming formulation for facility layout design using flexible bays

This paper presents a mixed-integer programming formulation to find optimal solutions for the block layout problem with unequal departmental areas arranged in flexible bays. The nonlinear department area constraints are modeled in a continuous plane without using any surrogate constraints. The formulation is extensively tested on problems from the literature.

Unequal area flexible bay facility layout using ant colony optimisation

In this paper, an ant colony optimisation approach is proposed to solve the facility layout problem with unequal area departments and flexible bays, which is one of the commonly used layout representations in industry practice. Optimal approaches to the facility layout design can only solve problems with a limited number of departments. The proposed ant colony optimisation approach is tested on 21 fairly well-known unequal area facility layout problems from the literature with up to 62 departments, and the results are compared with the previously best known solutions. Different cases of these problems are also tested. The proposed ant colony optimisation approach is shown to be very effective in finding previously known best solutions in a very short amount of CPU times and making improvements up to 17.38%.

Using Kolb's Experiential Learning Cycle to improve student learning in virtual computer laboratories

In information security education, learning experiences that involve hands-on experimentation are extremely important. However, information security topics are challenging to teach in traditional computer laboratories mainly due to restrictive information technology policies. In the literature, virtual computer laboratories have been proposed to address the challenges of providing students with hands-on learning experiences in information security. While the literature mainly focuses on technical aspects of virtual computer laboratories and related hands-on activities, pedagogical aspects of hands-on activities are overlooked. Our experiences with a virtual computer laboratory have shown that hands-on activities which are designed based on a prescriptive, step-by-step approach do not always achieve the expected learning outcomes. In this paper, we propose Kolbs Experiential Learning Cycle as a framework to design hands-on activities in virtual computer laboratories, and we argue that hands-on activities designed based on this framework enhance student learning outcomes. We illustrate how the stages of Kolbs model can be incorporated into hands-on activities and present results from two empirical studies to test the effectiveness of the proposed framework. The empirical findings in the first study suggest that hands-on activities designed based on the proposed framework are more likely to increase student interest and competency compared to step-by-step hands-on activities. In the second study, the collected data is analyzed using structural equation modeling to determine the relationships among the factors affecting student learning outcomes as a result of hands-on activities. The results of the second study show that student-to-student interaction is an important factor determining student learning experiences. Hands-on learning is studied in virtual computer laboratories.Cookbook activities do not achieve comprehensive learning.To enhance student learning, Kolbs model is proposed to design hands-on activities.The benefits of Kolbs model are investigated in the information security domain.The proposed approach enhances student learning in virtual computer laboratories.

A new relaxed flexible bay structure representation and particle swarm optimization for the unequal area facility layout problem

The facility layout problem (FLP) with unequal area departments is a very hard problem to be optimally solved. In this article, a hybrid particle swarm optimization (PSO) and local search approach is proposed to solve the FLP with unequal area departments. The flexible bay structure (FBS), which is a very common layout in manufacturing and retail facilities, is used. Furthermore, the FBS is relaxed by allowing empty spaces in bays, which results in more flexibility while assigning departments in bays. The proposed PSO approach is used to solve the FLP instances from the literature with varying sizes. The comparative results show that the PSO approach is very promising and able to find the previously known-optimal solutions in very short CPU times. In addition, new best solutions have been found for some test problems. Improvements have been achieved by allowing partially filled bays.

A hybrid genetic algorithm approach for backbone design of communication networks

The paper presents a hybrid approach of a genetic algorithm (GA) and local search algorithms for the backbone design of communication networks. The backbone network design problem is defined as finding the network topology minimizing the design/operating cost of a network under performance and survivability considerations. This problem is known to be NP-hard. In the hybrid approach, the local search algorithm efficiently improves the solutions in the population by using domain-specific information while the GA recombines good solutions in order to investigate different regions of the solution space. The results of the test problems show that the hybrid methodology improves upon previous approaches.

Linear Programming Based Genetic Algorithm for the Unequal Area Facility Layout Problem

The facility layout problem (FLP) is generally defined as locating a set of departments in a facility with a given dimension. In this paper, a hybrid genetic algorithm (GA)/linear programming (LP) approach is proposed to solve the FLP on the continuous plane with unequal area departments. This version of the FLP is very difficult to solve optimally due to the large number of binary decision variables in mixed integer programming (MIP) models as well as the lack of tight lower bounds. In this paper, a new encoding scheme, called the location/shape representation, is developed to represent layouts in a GA. This encoding scheme represents relative department positions in the facility based on the centroids and orientations of departments. Once relative department positions are set by the GA, actual department locations and shapes are determined by solving an LP problem. Finally, the output of the LP solution is incorporated into the encoding scheme of the GA. Numerical results are provided for test problems with varying sizes and department shape constraints. The proposed approach is able to either improve on or find the previously best known solutions of several test problems.