DSM models/Energy/SH In a smart grid

 Subproblem-1  : 

Load Scheduling 

for Islanded SH In a smart grid, the DSM reduces the energy consumption cost on the demand side and reduces energy production loads on the supply side. Researchers have proposed many DSM models, infrastructures, and platforms to optimize power demands. In the distributed DR, the supply side broadcasts the information, e.g., current load and tariff. Individual consumers optimize energy consumption [62]. However, error messages on the communication network and locally solving the optimization are challenges for efficient power management. The research [33] proposes a cloud-based centralized DSM. To overcome the delay issue of the cloud, the researchers introduce a two-tier cloud-based model: edge cloud near the end-users and centralized cloud for all edge clouds. Inefficient resource usage of cloud computing 11 exacerbates the delay issue. Load traffic on cloud-based DSM and network delay are the challenges for DSM in SG [63]-[66]. To overcome the delay issues, the research [55] proposes a fog-based architecture for DSM. The fog has limited resources compared to the cloud;

however, it is close to the end devices. A bottleneck is created for fog computing due to a huge load of requests that compromise the performance of fog computing. Routing of the demand-side requests’ traffic at the appropriate data center and allocation of requests on fog computing resources for enhanced performance are challenging. A system model and solutions have been proposed in Section 3.4 to address these challenges. The simulations and results in Section 4.2 validate the efficiency of solutions. 

1.6.2 Subproblem-2:

 Efficient Resource Allocation Techniques in Fog The authors in [67] propose an energy-efficient layered architecture for cloud and fog-based systems. However, task management for efficient RT and PT is essential from a conconsumer's perspective. The optimized execution cost requires energy-efficient resource utilization of high-performance infrastructure. The research [68] proposes the weighted Round Robin (RR) as SBP for the selection of potential data center and task allocation to the resources are performed with a honey bee scheme. However, the research [68] proposes priority-based task allocation to VMs using the formulated objective function of a honey bee algorithm. This algorithm has latency issues, which leads to the starvation of low-priority tasks. However, Munir et al. [69] propose an architecture for IoT to optimize service time for end devices. The combination of efficient SBP and load balancing algorithm are required to optimize the service PT and computing cost with a different load of requests, and a variety of physical and virtual resources. In Section 3.5, the system model is proposed in which a solution for efficient SBP and load balancing algorithm to optimize PT and computing cost is analyzed. The simulations and results validate the performance of the proposed SBP and load-balancing algorithm in Section 4.3.

 1.6.3 Subproblem-3: Real-time Energy Management Service for Community

In SG, energy crises have been a serious concern for many years. The shortfall of power generation to fulfill the demand and pollution of the environment during the process of energy generation are the energy crises. The researchers are addressing such issues by proposing energy management techniques using load shifting and dynamic pricing schemes[70]. The 12 research[71] extends the efficient energy management and reliability of HEMS [70]. Towards the improvement on [71], in [72], the authors proposed an infrastructure to allocate cloud resources with flexibility to optimize cost for DSM in SHs. However, delayed responses from the cloud are not suitable for delay-sensitive applications. Thus, authors in [59], [73], [74] support fog computing over the cloud because of increasing demands. VMs are installed in computing environments for efficient resource utilization and processing. In [75], experimental analysis is conducted to evaluate the number of VMs for a physical machine in an energy-efficient way. However, the authors concluded that the capacity of VMs and their collocation reduce the job completion time. Appropriate resource allocation techniques, size, and number of VMs for physical resources are important for cost-efficient and real-time energy management service(s). In Section 3.6, a system model is proposed in which very large communities are provided with power management services from their respective fog. Two types of load-balancing algorithms are used to allocate the load of requests on VMs. The simulations are performed with three scenarios and the results are discussed in Section 4.4, the simulation results validate the efficiency of fog-based energy management service as compared to the cloud. The effects of too many VMs and types of load-balancing techniques on service execution time, computing cost, and RT for power users are studied. 

1.6.4 Subproblem-4: Real-time Energy Management Service for Communities of Prosumers 

The energy consumption optimization by shifting load from on-peak to off-peak time by individual SH of a community creates rebound peaks in off-peak time and increases userdiscomfort [76]. The integration of RESs and ESSs shaves on-peak load demand and reduces user discomfort [77]. The integration of RESs with the local utility or power grid can reduce energy consumption costs and reduce traditional fossil fuel-based power generation; however, it is challenging due to the intermittency and uncertainties of RESs [78]. The cloud-asan-energy management is proposed for prosumers and energy-producing companies [61]. The authors extended their contribution by introducing Electrical Vehicles (EVs) with fair pricing in the existing system [31]. However, the authors agree that the computation time of the cloud increases with the increase in the number of prosumers. The computing time for 1000 prosumers is 7 seconds and 3.2 minutes for 15,000 prosumers. The RT affects the power trade and cost, which is not considered in this work. Moreover, the cloud-based system has a long delay for response and it increases with long PT [41]. Promoting the integration of 13 of DERs and power grids with near real-time energy management services is challenging. In Section 3.8, three power sources (large scale, small scale, and utility) are integrated to optimize power cost for the community by following the proposed contracts. The appropriate choice of physical and virtual resources along with efficient resource utilization techniques are performed. The simulation results in Section 4.6 validate the optimized cost for energy utilization by the integration of renewable energy with near real-time service.

 1.6.5 Subproblem-5: Real-time Energy Management Service for Community of Prosumers and Consumers 

In the second wave of restructuring the electricity market of the United States during 2000- 2001, the prime focus was economic benefits, which lead to new developments of SGs. These developments lead to the third reform, in which the demand side considers RESs, ESSs, and convergence of distributed energy sources [79]. The governments launched subsidized RESs projects to control environmental issues [80]; however later, these projects were stopped due to ignoring the unwillingness to pay back [81]. In the literature, researchers propose business-oriented distributed and centralized renewable energy-sharing models and techniques [82], [83] to minimize energy costs for consumers of communities. However, the authors provide a solution for either the end-user side or service providing side. Chen et al. [61], propose a cloud-based energy management system for prosumers; however, the computing side is not discussed in detail including computing resources, RT, computing cost, and energy consumed by the computing system to provide services. Moreover, the authors admit that the longer PT with an increased number of prosumers increases the RT and degrades the power system of the community. Hence, real-time service for power prosumers and consumers to promote high integration of renewable energy with cost and energy-efficient computation is challenging. In Section 3.9, a system model to integrate utility and prosumers’ power honoring the contract is proposed. A Linear Programming (LP) model is proposed, which provides an incentive to the community with a cheap power supply and maximizes profit for the service provider. The service is provided using Fog’s computing resources. Efficient computation is performed with two combinations of SBPs and load-balancing algorithms. The simulations and results in Section 4.7 validate the cost-efficient power utilization for the community with the proposed LP model. The PT, RT, and energy consumption with the proposed combination of SBP and load balancer are efficient.