1. Introduction

2. Literature review

3. Problem definition

4. Methodology

5. Validation experiments & results

6. Case study: “the Green River in the new administrative capital in Egypt”

7. Conclusions

References

Abstract

Surveillance cameras have a plethora of usages in newly born cities including smart traffic, healthcare, monitoring, and meeting security needs. One of the most famous new cites is the Egypt’s new administration capital “New Cairo”. The new administration capital of Egypt mainly characterizes with the green life style via the “Green River “. In this paper, a new enhanced Artificial Intelligence (AI) algorithm is introduced for adjusting the orientation of Pan–Tilt–Zoom (PTZ) surveillance cameras in new Cairo. In other words, the new proposed algorithm is used for improving the field of view (FOV) coverage of PTZ cameras network. For validating the proposed algorithm, it is tested on many scenarios with different criterions. After that, the proposed algorithm is applied to adjust the PTZ monitoring cameras in the green river which locates on new administrative capital as an equivalent to the river Nile. In addition, it compared with several other AI algorithms through the appropriate statistical analysis. The overall experimental results indicate the prosperity of the proposed algorithm for increasing the coverage of the PTZ surveillance system.

Introduction

The wave of technological development has swept through our world until most of the modern residential cities have become smart cities characterized by automation and intelligence. The smart city can be defined as a high-level ontology that describes the semantic categories including a series of innovations in urban systems sustained by broadband networks, sensors, data management technologies, software applications and e-services (Komninos & Mora, 2018). Roughly speaking, the smartness of the city is determined by the Structure and Functions of its Semiotics (Data, Information, Knowledge) management system. So, these functions can be summarized as (Ramaprasad, Sánchez-Ortiz, & Syn, 2017):

Therefore, one of the most important pillars of such a smart urban is the surveillance systems so-called image sensors/camera networks, which can be considered vision organs of the smart city.

Network cameras are used in smart cities across the globe for monitoring, surveillance and other security needs. Due to the Internet of Things (IoT) comes to life; they extend well beyond simple surveillance tasks to other application scenarios. There are several types of cameras which differ in servo capabilities, sensor element, lens type, etc. In general, they can be categorized into three main types (Erdem & Sclaroff, 2006): Fixed Perspective, Omni-directional, and Pan-TiltZoom (PTZ) cameras. The former has a fixed position, orientation, and focal length. The second has a full 2 horizontal coverage but it has a small focal length which may cause undesirable lens aberration effects. The last type of cameras (PTZ) is adjustable in a predefined range. They can be rotated horizontally and vertically around their (Tilt) and (Pan) axis respectively. Also, some PTZ cameras have an adjustable focal length (Zoom).

This paper aims at scheduling bag-of-tasks (BoT) applications under budget constraints on hybrid clouds for minimizing the makespan. To solve this NP-hard problem, we propose a novel firefly algorithm (NFA) in which the evaluation of a firefly consists of two steps: (1) mapping a firefly to a scheduling solution (a task sequence); (2) calculating the solution’s objective (its corresponding makespan). In the first step, different from the well-known ranked-order value (ROV) rule, we propose a distancebased mapping operator that relies on the distance between a firefly and the brightest one to determine the mapping relationship between a firefly and a solution. We use a probability model in which solutions corresponding to fireflies closer to the brightest one would have higher probabilities to inherit tasks from the current best solution. In this manner, these solutions can inherit more ‘‘good genes’’ hidden inside the current best solution to evolve into more high-quality solutions. In the second step, we employ an effective heuristic to evaluate solution objectives. We further develop a composite heuristic to generate the initial best solution, providing the proposed NFA with a good start. We also establish a new movement scheme such that fireflies distant from the brightest one can explore a wide range in the search space, whereas fireflies nearby the brightest one can search in a small neighborhood. Experimental results show that, by employing the above-mentioned strategies, NFA outperforms the standard firefly algorithm and the existing best algorithm, in terms of scheduling effectiveness and computational efficiency. Specifically, the distance-based mapping operator is verified to be both more effective and more efficient than the ROV rule. The composite heuristic is capable of generating a good initial solution, leading to the high quality of the final schedule. The movement scheme can further reduce the makespan of BoT applications.

Introduction

Cloud computing is a popular distributed computing paradigm that can deliver a massive amount of computing resources in a pay-as-you-go manner. In other words, users only need to pay what they have actually used. As a result, users can temporally use computing resources of public clouds while the private cloud cannot provide sufficient resources. In order to achieve this target, cloud vendors offer hybrid cloud solutions that integrates a private cloud with public clouds seamlessly. As illustrated in Figure 1, with the aid of hybrid cloud solutions, administrators or programs of the private cloud can manage the computing resources of the hybrid cloud (i.e., the private cloud and public clouds) via unified interfaces. For example, in case that the available computing resources of the private cloud cannot afford an instance of a ‘‘large’’ virtual machine (VM) type for processing a specific task, an instance of the same VM type can be created on a public cloud to tackle this task. In addition, many cloud providers now can deliver and charge VM instances in seconds (e.g., QingCloud1 and TecentCloud2 ) instead of in hours. This new charging mechanism makes hybrid clouds more popular for cloud customers, since they would not waste a certain fraction of the last hour that has to be paid entirely in traditional charging mechanisms.

1- Introduction

2- Related work

3- Problem origination

4- Proposed approach

5- Experiment and evaluation

6- Strength and weakness of the proposed method

7- Conclusion and future scope

References

Abstract

Nowadays, with the widespread access to web 2.0, the social network plays an unbelievable role in knowledge sharing and diffusion of new products. People can share their views and can visit other’s opinion about the particular material, news, products, artifacts and, trends, etc. anywhere, anytime, and anywhere. An Opinion leader is a critical person who can change, modify and transform other’s view by their knowledge and proficiency. In this article, an innovative approach is proposed to discover the top-N local and global opinion leader within the community and social network respectively. Initially, we identified the community structure within the social network using the modified Louvain method and next identified the opinion leader using a modified firefly algorithm in each community. We also determined the global opinion leader within the same social network using the same firefly algorithm. The proposed approach is exceptionally supportive to expert and intelligent system because it competently discovered the local optimum concurrently in each subgroup of the social network. All the users can update its attractiveness value without any supposition, and as soon as the distance among the user’s increases, the other users can automatically create another subgroup in the network and form the local community. In addition, as the population size in the network increases, the entire users measure their prominence simultaneously. Therefore, there is no consequence on computational time and accuracy of the algorithm. Thus, the proposed algorithm is superlative suitable for discovering the opinion leader in the local community and globally in the social network. For legalized the proposed approach, we implemented our proposed method on synthesized as well as on real dataset. Finally, we concluded that both the recommended procedures are much better concerning the accuracy, precision, recall, and F1-score with the widely used standard Social Network Analysis (SNA) measures.

Introduction

In the current web scenario, social networking is an integral part of human life. These sites, such as Facebook, Twitter, Tumblr, and Instagram, etc. provide the opportunity to interact with another unknown world of known things and human being. Currently, these Websites are the primary source of information transmission and dissimilation (Boyd & Ellison, 2007; Scott, 2000; Wasserman & Faust, 1994). Moreover, social networking sites often provide a platform for the companies for the diffusion of the new product and merchandise (Abrahamson & Rosenkopf, 1997; Bakshy, Rosenn, Marlow, & Adamic, 2012). The success rate of this diffusion is dependent upon identifying the critical users effectively in the social network, such type of person known as an opinion leader. Opinion leader has more incredible significance for the diffusion of the new product. Opinion leader also affects the consumer behavior and decision making by their knowledge and experience about a particular product (Chan & Misra, 1990; Myers & Robertson, 1972). According to Dye (2000), an opinion leader is a person or set of persons having more influence on the customer’s adoption process and decision making. Lazarsfeld et al. have introduced the phenomenon of the opinion leader in their seminar in 1940 and 1950 (Gold, Katz, Lazarsfeld, & Roper, 1956). According to them, identification of opinion leader is a two-step procedure in which in the first step, opinion leader analyzes, examine and understand the end user’s requirements, and in the second step, opinion leader derives their own opinion from the first step incorporated with their knowledge and skills.

1- Introduction

2- Related work

3- Proposed work

4- Comparative study

5- Conclusion

References

Abstract

The Firefly algorithm is a population-based optimization algorithm. It has become popular in the field of optimization and has been applied to engineering practices. Recent works have failed to address how to find the global minimum because their algorithm was trapped in the local minimum. Also, they were not able to provide a balance between exploration and exploitation. In this paper, the Tidal Force formula has been applied to modify the Firefly algorithm, which describes the effect of a massive body that gravitationally affects another massive body. The proposed algorithm brings a new strategy into the optimization field. It is applied by using exploitation (Tidal Force) and keeping a balance between the exploration and exploitation on function suitability. Plate shaped, Steep Ridges, Unimodal and Multimodal benchmark functions were used to compare experimental results. The study findings indicate that the Tidal Force Firefly algorithm outperforms the other existing modified Firefly algorithms.

Introduction

Population-based optimization techniques have become widespread in the last two decades. Optimization problems are examined in a variety of fields, which consist of highly nonlinear, multimodal, multidimensional, and differentiable functions. However, traditional optimization techniques have not been able to solve these problems. The population-based techniques, with their own robustness and flexible behavior in solving optimization problems, bring novel insights in order to solve the problems instead of using traditional optimization techniques. The PSO algorithm is inspired by birds’ behavior [1], which defines that all birds move towards the best bird. The PSO works with two populations, such as best position and current positions. Diversity solutions are one of the advantages of the PSO, which performs better than the single point algorithms. An ant colony is another popular algorithm that was designed for optimization problems. The algorithm is based on the behavior of ants and was proposed by Dorigo. Ants search for a best path solution between their colony and a food source. Each ant randomly moves towards a destination ant. The paths are followed by ants, based on the probability of pheromones [2]. Differential evolution based on individual’s differences (called the DE algorithm) is similar to the GA algorithm that defines specified crossover, mutation and selection [3]. It computes parallels and takes the best result in a ∗ Corresponding author. E-mail addresses: arefyelghi@ktu.edu.tr (A. Yelghi), ckose@ktu.edu.tr (C. Köse). few dimensions. The Harmony Search (HS) algorithm was modeled by taking inspiration from the improvisation of musicians [4]. The algorithm uses musical Pitch Range, Harmony, Aesthetics, Practice and experience in the algorithm, which links to decision variables, iteration concepts and so on. The harmony (solution) is produced randomly and checks with stored solutions to place better solutions in their place.