Implementation of roulette wheel and random selection methods in a hybrid intelligent system: A comparison study for two Islands and Subway distributions considering different router replacement methods
Introduction
During the last couple of decades there has been an unprecedented growth of wireless networks and Artificial Intelligence (AI) applications. This trend is expected to continue for the foreseeable future [1], [2], [3], [4]. Especially, with the appearance of Internet of Things (IoT), different kind of networks and applications are appearing and they support us in everyday life. The Wireless Mesh Networks (WMNs) have received much interest and attention because of their self-organizing and self-configuring abilities. The nodes of WMNs automatically initiate and maintain the mesh connectivity, thus reducing the overall costs and enabling efficient network maintenance, robustness, and good coverage [5]. The WMNs have good characteristics and can be used in different scenarios and areas with difficult terrains. They can be deployed as enterprise and community networks, in addition to applications for cities, intelligent transportation systems and surveillance systems.
Designing and engineering WMNs is not an easy task because this process includes many parameters which are often uncorrelated and therefore a trade-off on the solutions is required in the optimization process. The parameters include mesh routers connectivity, coverage of mesh clients, Quality of Service (QoS), and the network cost. The first three parameters should be optimized while keeping the overall cost of the network low.
Because there are more than two uncorrelated parameters for optimization, the problem becomes NP-hard, which is a problem that cannot be solved in polynomial time [6], [7], [8]. Recently, there are different approaches such as machine learning, AI, soft computing and evolutionary-based algorithms which are capable of solving NP-hard problems [9], [10], [11].
There are many intelligent algorithms applied to the networking field such as Particle Swarm Optimization (PSO) [12], Genetic Algorithms (GAs) [13], Tabu Search (TS) [14], Simulated Annealing (SA) [15], Hill Climbing (HC), and Neural Networks (NNs). In literature there are also many optimization algorithms such as Harris Hawks Optimization (HHO) [16], Coyote Optimization Algorithm (COA) [17], Grey Wolf Optimizer (GWO) [18], Ant Colony Optimization (ACO) [19], Ant Lion Optimizer (ALO) [20], Whale Optimization Algorithm (WOA) [21], Marine Predators Algorithm (MPA) [22], Slime Mould Algorithm (SMA) [23], Sine Cosine Algorithm (SCA) [24], Reptile Search Algorithm (RSA) [25], Coronavirus Herd Immunity Optimizer (CHIO) [26], Ebola Optimization Search Algorithm (EOSA) [27], African Vulture Optimization Algorithm (AVOA) [28], Aquila Optimizer (AO) [29], Bald Eagle Search optimization (BES) [30], Artificial Bee Colony (ABC) [31], Glowworm Swarm Optimization (GSO) [32], Firefly Algorithm (FA) [33], Bat Algorithm (BA) [34], and Salp Swarm Algorithm (SSA) [35].
These meta-heuristic optimization algorithms are nature- inspired computational intelligence solutions, which have high performance and can find optimal solutions for complex real-life problems in engineering and other scientific works. They can be grouped in evolutionary, swarm-based, physics-based and human-based. They have different characteristics and can be applied in different fields considering the advantages and disadvantages.
In our previous work, we implemented some simulation systems for WMNs considering simple meta-heuristic algorithms such as PSO-based simulation system called WMN-PSO system, HC-based simulation system called WMN-HC, and GA-based simulation system called WMN-GA. Considering the results from these research works and the advantages and disadvantages of implemented systems, we improved our research by designing and implementing hybrid intelligent simulation systems based on PSO, HC and Distributed GA (DGA), which were named WMN-PSOHC, WMN-PSODGA, and WMN-PSOHC-DGA, respectively. In these research works, for the fitness function, we consider two parameters the Size of Giant Component (SGC) and the Number of Covered Mesh Clients (NCMC).
In this research work, different from the previous systems, for WMN-PSODGA simulation system we consider another parameter for optimization, called NCMCpR, which shows the Number of Covered Mesh Clients per Router. By considering this parameter, we optimize the load between mesh routers in a WMN. Our implemented WMN-PSODGA simulation system with load balancing can provide high network connectivity, full client coverage, and good load balance between mesh routers.
We consider different router replacement methods for each distribution of mesh clients in order to find good solutions for different scenarios. In our previous work [36], we have considered Normal, Weibull, and Boulevard distributions of mesh clients and have presented the results for five router replacement methods. In this paper, we consider two new distributions, such as Two Islands and Subway distributions, and present a comparison study not only for different router replacement methods but also for roulette wheel and random selection methods.
The contribution of this work are as follows.
- •
We improve WMN-PSODGA simulation system by considering the load balancing between mesh routers and by implementing Two Islands and Subway distributions of mesh clients.
- •
We present a comparison study not only for different router replacement methods but also for roulette wheel and random selection methods.
- •
We consider different router replacement methods for each distribution of mesh clients in order to find good solutions for different real environment scenarios.
The remainder of the paper is organized as follows. Section 2 presents intelligent algorithms and a review of relevant works. Section 3 introduces the WMN-PSODGA system. Section 4, discusses the simulation results. Section 5 summarizes the findings of the paper and gives thoughts over future work.
Section snippets
Background and related work
In this section, we introduce the intelligent algorithms used in our hybrid intelligent system and review several research papers to show how our system is different from works existing in the literature.
WMN-PSODGA hybrid intelligent system
We show the overview of the proposed WMN-PSODGA hybrid intelligent system in Fig. 3, its flowchart in Fig. 4, and its pseudocode in Algorithm 3. In WMN-PSODGA system there are two parts: the PSO part and the DGA part. In the PSO part, we have particle patterns, while in the DGA part, there are islands and individuals. The individuals in DGA are swapped between islands. Each parts provides its own solution and in each iteration some of them are exchanged using a migration function (see Fig. 4).
Simulation results
In this section, we present the results of Two Islands and Subway distributions of mesh clients for different router replacement methods using the roulette wheel and random selection methods. The simulation setup of WMN-PSODGA hybrid intelligent system is given in Table 2.
Conclusions
WMNs are a low-cost alternative to other wireless networks, which require the deployment of expensive infrastructure, but they face problems of their own kind, too. One of these problems is the placement of mesh routers over an area of interest. To keep down the costs but without compromising the network performance, the mesh routers should be placed in positions that have them as interconnected as possible while still offering services to every client. Moreover, to deliver the services in the
CRediT authorship contribution statement
Admir Barolli: Conceptualization, Methodology, Formal analysis. Kevin Bylykbashi: Software, Investigation, Writing – original draft, Writing – review & editing. Ermioni Qafzezi: Writing – original draft, Data curation. Shinji Sakamoto: Software, Visualization. Leonard Barolli: Validation, Supervision, Writing – review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (50)
- et al.
Wireless mesh networks: A survey
Comput. Netw.
(2005) - et al.
Harris hawks optimization: Algorithm and applications
Future Gener. Comput. Syst.
(2019) - et al.
Grey wolf optimizer
Adv. Eng. Softw.
(2014) The ant lion optimizer
Adv. Eng. Softw.
(2015)- et al.
The whale optimization algorithm
Adv. Eng. Softw.
(2016) - et al.
Marine predators algorithm: A nature-inspired metaheuristic
Expert Syst. Appl.
(2020) - et al.
Slime mould algorithm: A new method for stochastic optimization
Future Gener. Comput. Syst.
(2020) SCA: A Sine Cosine Algorithm for solving optimization problems
Knowl.-Based Syst.
(2016)- et al.
Reptile Search Algorithm (RSA): A nature-inspired meta-heuristic optimizer
Expert Syst. Appl.
(2022) - et al.
African vultures optimization algorithm: A new nature-inspired metaheuristic algorithm for global optimization problems
Comput. Ind. Eng.
(2021)