Figure 1 - uploaded by Junichi Suzuki
Content may be subject to copyright.
Source publication
This paper proposes and evaluates an evolutionary multiobjective optimization algorithm, called EVOLT, which heuristically optimizes quality of service (QoS) parameters in communication networks. EVOLT uses a population of individuals, each of which represents a set of QoS parameters, and evolves the individuals via genetic
operators such as crosso...
Contexts in source publication
Context 1
... power delivery system transmits generated electricity from power stations to consumers (Fig- ure 1) [29,30]. Electricity is transmitted in high voltage (e.g., 110 KV) from a power station in order to reduce energy loss in transmission, and distributed toward consumers through a chain of substations. ...
Context 2
... most of substations and some of power stations are unmanned, power utilities remotely monitor and control them with communication networks [31]. Figure 1 shows an example network that consists of a control center, substations and a power station. Each substation and power station periodically monitors its operations and equipment (e.g., every few seconds), and transmits the monitored status information to a control center. ...
Context 3
... power utility communication network is often configured as a tree structure where a control center serves as its root (Figures 1 and 2). This paper considers two types of IP networks: a smaller-scale network of 34 nodes (a control center, 30 substations and 3 power stations) and a larger-scale network of 67 nodes (a control center, 60 substations and 6 power stations). ...
Context 4
... 9: Fitness-based Crossover Operator Figure 10 shows an example on how to determine offspring's QoS parameters. This exam- ple assumes that the second parent (p 2 ) has a higher fitness value than the first parent (p 1 ): The proposed crossover operator is designed following a property in Holland's schema theo- rem [34,35], which proves that one-point crossover contributes to improve the average fitness values of individuals through generations. ...
Context 5
... this property, Holland's schema theorem assumes that offspring are placed either inside or outside the region bounded by their parents. The proposed crossover operator emulates this property as shown in Figure 10; offspring's QoS parameters are placed as either ...
Context 6
... C(NSGA-II, EVOLT ) are 97% and 0%, respectively, as discussed in Section 4.2 ( Table 7). Figure 11 compares the optimality of EVOLT 's variations with the generation distance (GD) ...
Context 7
... is the number of objectives. As Figure 11 illustrates, EVOLT variations converge individuals generation by generation. ...
Context 8
... means that EVOLT converges approximately two times faster than NSGA-II. Figure 11 demonstrate that EVOLT 's operators complement with each other well and their combination successfully balances the optimality, diversity (i.e., distribution and spread) and convergence speed of individuals while satisfying given QoS requirements. ...
Context 9
... maps each solution, which is high-dimensional (six dimensional) data, on a low-dimensional (two dimensional) space. For example, Figure 12 shows a 25×25 SOM that maps 100 solutions obtained from a particular simulation with a smaller-scale network. (All individuals are constraint-dominated at the last generation.) ...
Citations
... The EAs have been used to find optimal solutions of various optimization problems in a wireless sensor network (WSN) [7]- [9]. A GA was proposed to optimize two objectives in [7]. ...
... The algorithm involves maximizing the coverage and the network lifetime. EVOLT was proposed in [9]. ...
... Next, constraint (8) indicates whether the total number of requests from client i to all service providers is greater than or equal to the number of requests from client i. Finally, constraint (9) indicates whether the number of requests and the capacity are positive values. However, it is very difficult to find the best value for the weights ω 1 , ω 2 , and ω 3 . ...
Currently, over nine billion things are connected in the Internet of Things (IoT). This number is expected to exceed 20 billion in the near future, and the number of things is quickly increasing, indicating that numerous data will be generated. It is necessary to build an infrastructure to manage the connected things. Cloud computing (CC) has become important in terms of analysis and data storage for IoT. In this paper, we consider a cloud broker, which is an intermediary in the infrastructure that manages the connected things in CC. We study an optimization problem for maximizing the profit of the broker while minimizing the response time of the request and the energy consumption. A multiobjective particle swarm optimization (MOPSO) is proposed to solve the problem. The performance of the proposed MOPSO is compared with that of a genetic algorithm and a random search algorithm. The results show that the MOPSO outperforms a well-known genetic algorithm for multiobjective optimization.
Participatory sensing enables to broadly collect information from mobile devices with embedded sensors. However, users have to use limited energy on the mobile devices as well as have to pay more communication fee when reporting sensory data to service providers. It is important to give them incentives to send information in participatory sensing systems. In this paper, we propose a novel incentive mechanism for participatory sensing based on the evolutionary algorithm. Our proposed algorithm is evaluated by extensive simulations and its performance is compared with state-of-the-art algorithms. The simulation results demonstrate that our algorithm successfully increases the number of participating users while it can also uniformly expand the coverage of a sensing area.
This paper proposes to apply a multiobjective optimization evolutionary algorithm in wireless sensor node redeployment process to improve network lifetime and sensing coverage. The multiobjective optimization uses a population of individuals, each of which represents a set of wireless sensor node positions, and evolves them via the genetic operations for seeking optimal sensing coverage and network lifetime. The data transmission success rate and the total moving cost are also added as constraints. Simulation results show that the proposed multiobjective optimization evolutionary algorithm outperforms a well-known existing evolutionary algorithm for multiobjective optimization.
This paper studies and evaluates a fitness-based crossover operator in an evolutionary multi-objective optimization algorithm, which heuristically optimizes the sensing coverage area and the installation cost in wireless sensor networks. The proposed evolutionary algorithm uses a population of individuals (or chromosomes), each of which represents a set of wireless sensor nodes' types and positions, and evolves them via the proposed fitness-based crossover operator (FBX) for seeking optimal sensing coverage and installation cost. Simulation results show that the fitness-based crossover evolutionary algorithm outperforms a well-known existing evolutionary algorithm for multi-objective optimization.
