Article Preview
TopIntroduction
In the age of information, the continuous demand for powerful systems that can provide rapid data collection with the minimum price is attracted the attention of both academic and industrial communities. This interest triggers remarkable technological development that paved the way for the development of a successful solution in data gathering called Wireless Sensor Network (WSN) (Rostami et al., 2018; Khan et al., 2018). The WSN is an assemblage of spatially dispersed sensors that varies from a few to thousands (Kumar et al., 2019). Each sensor has the capability to sense the surroundings and then communicate the gathered data with its neighbors until it reaches the end-user (Sangwan et al., 2015). In general, wireless sensors are constrained in terms of resources, energy, and sensing range (Tiegang et al., 2014). These limitations led to the emergence of several complex problems related to the design of WSNs. One important problem that affects many intrinsic performance criterions such as coverage, cost, and lifetime is the deployment optimization problem (Tsai et al., 2015). The complexity of the deployment problem arises from the constrained nature of sensors and the different application requirements. So there is no deployment technique capable to satisfy all usage requirements as several objectives are usually conflicting (Fellah et al., 2017). One common objective between many applications, which is considered the main purpose behind the usage of WSNs, is coverage (Sharma et al., 2019). Depending on the WSN's usage, whether it is for point surveillance, target tracking, or area surveillance, the goal is to place the sensors in optimal locations to achieve the maximum possible coverage and ensure a complete collection of data (Priyadarshi et al., 2020). The coverage reflects the capability of sensors to monitor the events in the sensing region (Deng et al., 2019). Therefore, coverage maximization is an essential task to be considered while placing the sensors in the monitoring area (Abdollahzadeh et al., 2016). In addition to coverage, many applications aim to achieve the maximum possible coverage with the least number of used sensors. By deploying a smaller number of sensors, several problems can be avoided, such as coverage redundancy and communication overhead. Furthermore, the deployment of a smaller number of sensors reduces the WSN’s deployment cost because deploying fewer sensors decreases the monetary expenditure of the user.
Achieving multiple design objectives, such as maximum coverage with the minimum cost puts the deployment optimization problem in the category of the NP-hard problems. Besides, optimizing conflicting objectives adds an additional layer of complexity to the deployment problem, which makes the choice of the proper deployment technique a very critical task. Metaheuristic algorithms as an alternative to exact methods have been used for obtaining the optimal solutions to various engineering design optimization problems (Mirjalili, 2015). The reason behind using metaheuristics lies in their advantages over exact methods that require a lot of computational time to achieve a satisfactory solution (Saremi et al., 2017). In this paper, a new version of the Black Widow Optimization (BWO) algorithm called Enhanced BWO (EBWO) is proposed to solve the deployment optimization problem in WSNs. The EBWO tries to determine the lowest sensor count and the best locations to deploy them to satisfy both coverage and deployment cost requirements. The main enhancements presented in this paper are as follows: