Salman A. Khan

A fuzzy particle swarm optimization algorithm for computer communication network topology design

Particle swarm optimization (PSO) is a powerful optimization technique that has been applied to solve a number of complex optimization problems. One such optimization problem is topology design of distributed local area networks (DLANs). The problem is defined as a multi-objective optimization problem requiring simultaneous optimization of monetary cost, average network delay, hop count between communicating nodes, and reliability under a set of constraints. This paper presents a multi-objective particle swarm optimization algorithm to efficiently solve the DLAN topology design problem. Fuzzy logic is incorporated in the PSO algorithm to handle the multi-objective nature of the problem. Specifically, a recently proposed fuzzy aggregation operator, namely the unified And-Or operator (Khan and Engelbrecht in Inf. Sci. 177: 2692–2711, 2007), is used to aggregate the objectives. The proposed fuzzy PSO (FPSO) algorithm is empirically evaluated through a preliminary sensitivity analysis of the PSO parameters. FPSO is also compared with fuzzy simulated annealing and fuzzy ant colony optimization algorithms. Results suggest that the fuzzy PSO is a suitable algorithm for solving the DLAN topology design problem.

Topology design of switched enterprise networks using a fuzzy simulated evolution algorithm

Abstract The topology design of switched enterprise networks (SENs) is a hard constrained combinatorial optimization problem. The problem consists of deciding the number, types, and locations of the network active elements (hubs, switches, and routers), as well as the links and their capacities. Several conflicting objectives such as monetary cost, network delay, and maximum number of hops have to be optimized to achieve a desirable solution. Further, many of the desirable features of a network topology can best be expressed in linguistic terms, which is the basis of fuzzy logic. In this paper, we present an approach based on Simulated Evolution algorithm for the design of SEN topology. The overall cost function has been developed using fuzzy logic. Several variants of the algorithm are proposed and compared together via simulation and experimental results are provided.

Fuzzy Logic-Based Decision Making for Detecting Distributed Node Exhaustion Attacks in Wireless Sensor Networks

Recent advances in the field of wireless networks have entailed collateral growth in the number of possible malicious attacks against them. A significant amount of work has been done towards ensuring security of a class of such networks, namely, Wireless Sensor Networks. Considering the untrusted environments of operations of such networks, the threat of distributed attacks against constrained resources i.e. sensor power, computation and communication capabilities cannot be overlooked. In [1], we modeled a class of attack called a distributed denial of service attack in such networks, and proposed a pattern-based scheme to detect such attacks. The limitation of this proposed scheme was on the lack of a tradeoff mechanism between improved performance of the detection scheme (higher detection rates) and corresponding increase in the use of the energy resources of the sensor nodes participating in the detection process. In this paper, we propose a fuzzy logic-based approach towards achieving demarkation in the values of specific parameters of the detection scheme, so as to ascertain a reasonable tradeoff between attack detection and node energy utilization. Simulation results depict the use of a fuzzy-based approach for addressing the energy-detection rate tradeoff problem effectively.

Computational intelligence techniques for placement of wind turbines: A brief plan of research in Saudi Arabian perspective

Placement of wind turbines (WTP) in a wind farm is a complex optimization problem, consisting of a number of design objectives and constraints. Although computational intelligence techniques have been applied to solve different versions of this problem, use of these techniques has been very limited to date. In this paper, we identify a number of computational intelligence techniques that have not been fully explored, or not explored at all, to efficiently solve the WTP problem. A research plan to utilize computational intelligence techniques to wind farm design layout in the context of Saudi Arabia has been briefly discussed.

Fuzzy simulated evolution algorithm for topology design of campus networks

The topology design of campus networks is a hard constrained combinatorial optimization problem. It consists of deciding the number, type, and location of the active network elements (nodes) and links. This choice is dictated by physical and technological constraints and must optimize several objectives. Example of objectives are monetary cost, network delay, and hop count between communicating pairs. Furthermore, due to the nondeterministic nature of network traffic and other design parameters, the objective criteria are imprecise. Fuzzy logic provides a suitable mathematical framework in such a situation. We present an approach based on the simulated evolution algorithm for the design of campus network topology. The two main phases of the algorithm, namely, evaluation and allocation, have been fuzzified. To diversify the search, we have also incorporated tabu search-based characteristics in the allocation phase of the SE algorithm. This approach is then compared with the simulated annealing algorithm, which is another well-known heuristic. Results show that on all test cases the simulated evolution algorithm exhibits more intelligent search of the solution subspace and was able to find better solutions than simulated annealing.

Design and analysis of a fault tolerant hybrid mobile scheme

Mobile computing systems provide users with access to information regardless of their geographical location. In these systems, Mobile Support Stations (MSSs) play the role of providing reliable and uninterrupted communication and computing facilities to mobile hosts. The failure of a MSS can cause interruption of services provided by the mobile system. Two basic schemes for tolerating the failure of MSSs exist in the literature. The first scheme is based on the principle of checkpointing used in distributed systems. The second scheme is based on state information replication of mobile hosts in a number of secondary support stations. Depending on the replication scheme used, the second approach is further classified as a pessimistic or an optimistic technique. In this paper, we propose a hybrid scheme which combines the pessimistic and the optimistic replication schemes. In the proposed scheme, an attempt is made to strike a balance between the long delay caused by the pessimistic and the high memory requirements of the optimistic schemes. In order to find the best ratio between the number of pessimistic to the number of optimistic secondary stations in the proposed scheme, we used fuzzy logic. We also used simulation to compare the performance of the proposed scheme with those of the optimistic and the pessimistic schemes. Simulation results showed that the proposed scheme performs better than either schemes in terms of delay and memory requirements.

Fuzzy Evolutionary Hybrid Metaheuristic for Network Topology Design

Topology design of enterprise networks is a hard combinatorial optimization problem. It has numerous constraints, several objectives, and a very noisy solution space. Besides the NP-hard nature of this problem, many of the performance metrics of the network can only be estimated, given their dependence on many of the dynamic aspects of the network, e.g., routing and number and type of traffic sources. Further, many of the desirable features of a network topology can best be expressed in linguistic terms, which is the basis of fuzzy logic. In this paper, we present a fuzzy evolutionary hybrid metaheuristic for network topology design. This approach is dominance preserving and scales well with larger problem instances and a larger number of objective criteria. Experimental results are provided.

Hard Real-Time Task Scheduling in Cloud Computing Using an Adaptive Genetic Algorithm

In the Infrastructure-as-a-Service cloud computing model, virtualized computing resources in the form of virtual machines are provided over the Internet. A user can rent an arbitrary number of computing resources to meet their requirements, making cloud computing an attractive choice for executing real-time tasks. Economical task allocation and scheduling on a set of leased virtual machines is an important problem in the cloud computing environment. This paper proposes a greedy and a genetic algorithm with an adaptive selection of suitable crossover and mutation operations (named as AGA) to allocate and schedule real-time tasks with precedence constraint on heterogamous virtual machines. A comprehensive simulation study has been done to evaluate the performance of the proposed algorithms in terms of their solution quality and efficiency. The simulation results show that AGA outperforms the greedy algorithm and non-adaptive genetic algorithm in terms of solution quality.

An evolutionary algorithm for network topology design

The topology design of campus networks is a hard constrained combinatorial optimization problem, dictated by physical and technological constraints and must optimize several objectives. Furthermore, due to the non-deterministic nature of network traffic and other design parameters, the objective criteria are imprecise. Fuzzy logic provides a suitable mathematical framework in such a situation. We present an approach based on a simulated evolution algorithm for design of a campus network topology. Three variations of the algorithm are presented and compared. Results show that the third variation, namely, simulated evolution with tabu search characteristics gives the best result.

On the use of Unified And-Or fuzzy operator for distributed node exhaustion attack decision-making in wireless sensor networks

The effect of advances in the fields of ubiquitous computing, wireless communications and embedded system design has seen a corresponding rapid improvement of wireless sensor technology. Sensor networks have emerged as a platform for deployment and sustenance of critical applications that require real-time sensing and data acquisition for decision-making purposes. A significant number of malicious attacks against the security of such networks have been identified in recent times. Considering the untrusted environments of operations of such networks, the threat of distributed attacks against constrained sensory resources i.e. sensor power, computation and communication capabilities cannot be overlooked. In this paper, we propose a fuzzy logic-based approach towards achieving demarkation in the values of specific parameters of an attack detection scheme for detecting distributed node-exhaustion attacks in wireless sensor networks. Using the Unified And-Or (UAO) aggregation operator, we model and formulate a mechanism to achieve a tradeoff between frequent attack detection and sensor node energy utilization. Simulation results prove the effectiveness of our approach in addressing the issue of computing the optimal parameter values for achieving a reasonable tradeoff between attack detection rate and sensor node energy utilization rate.