Elsevier

Applied Soft Computing

Volume 74, January 2019, Pages 122-132
Applied Soft Computing

Genetic algorithm-based optimisation of load-balanced routing for AMI with wireless mesh networks

https://doi.org/10.1016/j.asoc.2018.10.003Get rights and content

Highlights

  • A new optimised load balanced algorithm for AMI networks.

  • It is proven for scalability with different count of smart meters.

  • The proposed method is more efficient than HWMP and Load aware-HWMP.

  • The method is compared with a recent state-of-the art method for the same AMI application.

Abstract

The advanced metering infrastructure (AMI) in a smart grid contains hardware, software, and other electronic components connected through a communication infrastructure. AMI transfers meter-reading data between a group of smart meters and a utility centre. Herein, a wireless mesh network (WMN) with a random mesh topology is used to deploy the AMI communication network. In a WMN, paths are identified using a hybrid wireless mesh routing protocol (HWMP) with a load balancing feature called load aware-HWMP (LA-HWMP). These paths reduce the demand on links with a minimal air time metric; however, the delay in the data transmission of certain smart meters is high, given the large number of retransmissions caused by packet drop. To avert this problem and enhance the end-to-end delay, a genetic algorithm is applied on the LA-HWMP to obtain the optimal path. The optimisation process will result in the selection of paths with minimal delay. The genetic algorithm is developed with a rank-based selection, a two-point crossover, and a random reset mutation with a repair function to eliminate duplicate entries. The proposed method is compared with the HWMP, the LA-HWMP, and a state-of-the-art method that uses a combination of the ant colony algorithm and simulated annealing (ACA-SA) for AMI networks of different sizes. The obtained results show that the path identified by the proposed method yields a shorter delay and higher throughput than paths identified using the other methods.

Introduction

Smart grid [1] is an enhanced form of the existing power grid [2] that ensures availability, reliability, and efficiency. AMI [3], [4] is a gateway where the grid integrates with the consumer. The smart grid communication infrastructure comprises three major communication infrastructures [5], including (i) home area networks (HANs): short-range communication that connects a number of devices and sensors, (ii) neighbourhood area networks (NANs): medium-range communication connecting HAN with WAN, and (iii) wide area networks (WANs) that connect many NANs to a central control system. The AMI is the key NAN component collecting smart meter data from the NAN and aggregates it before sending it to the central control centre. To implement the AMI, some basic requirements are essential [6], [7]. Coordinating all the functional requirements demands a suitable communication infrastructure [8]. The AMI requires an efficient two-way communication network to support smart meters that send periodic messages carrying meter reading information. The communication network must contain an efficient topology and a routing protocol. Yasin et al. [5] summarised various technologies for smart meter communication in smart grid communication infrastructure.

Several network topologies and protocols exist to construct an AMI communication network. Packet routing is an important aspect of AMI network operations. The routing method depends on the type of communication technology used to implement it. Herein, the AMI communication network is deployed using a mesh topology. Diegio et al. [9] examined all possible routing protocols belonging to assorted network technologies such as WMNs, WiMAX, and power line communication to deploy AMI communication. Miscellaneous AMI routing and scheduling methods are discussed in [10]. This paper represents the network as a graph comprising a set of vertices and edges. Studies [11], [12] indicate that the WMN is suitable to support medium to high range network. It is suitable to implement scalable, cost-effective, and self-healing networks in both rural and urban areas. HWMP is the default routing protocol for wireless mesh networks, drafted by the IEEE 802.11s standard [13]. In [14], the importance of the SINR in the feasible scheduling of the WMN is proposed. Each link is assigned a pre-determined link weight that is used for link activation. The link weight is used to schedule the routes and for routing. Optimising WMNs using routing protocols, such as OLSR and HWMP, with infrastructure architecture and ad-hoc architecture are discussed in [15]. The objective functions of WMN routing are to reduce end-to-end delay and increase throughput. Optimising routing and planning using a genetic algorithm for WMNs is explained in [16]. This paper describes the WMN scenario but does not discuss interference in WMNs.

Jing et al. [17] proposed a new optimisation algorithm called the ACA-SA that finds the optimal paths between smart meters and data collectors in the AMI with load balancing. It results in minimal delay and maximum throughput. The level of congestion is calculated using a load factor throughout the path. Two routing methods are proposed herein. The first is applied for delay-sensitive applications of AMI such as meter reading transmission and outage management, and the second is for AMI applications with low packet loss ratio such as pricing and on-demand response. The proposed GA-LA-HWMP method is compared with the routing method for delay-sensitive AMI applications. Yuan et al. [18] proposed a hybrid routing protocol for WMNs with load balancing. This method divides the load into three levels: a mesh client-level load, a mesh router-level load, and a gateway-level load. The load is balanced within each level, and the results show that the packet loss is high in case of a large number of simultaneous flows.

Sahasrabuddhe and Mukherjee [19] formulated the multicast problem as an optimisation problem with the objective of minimising the average hop distance and number of transceivers. This method reduces the effects of interference. A multi-objective optimisation of WMNs to improve QoS parameters is discussed in [20]. The algorithm contains multiple objectives to enhance the QoS parameters using the ant colony routing method. In [21], AMI is deployed with a WMN that uses Zigbee and IEEE 802.11 technologies. Different routing methods and routing parameters are compared. This paper discusses hybrid network scenarios for AMI with both technologies. In [22], a multipopulation parallel genetic algorithm is used. This paper combines a GA with a simulated annealing technique, and details pertaining to the GA’s application in the deployment of inter-urban mesh networks are explained. In this work, optimisation enhances the routing method to satisfy the standard of IEEE 802.11s for the WMN. In [23], the optimisation of multicast request handling is implemented using a genetic routing algorithm. A number of alternate multicast paths are used to find the optimal path. The objective of the work is to minimise the number of split-capable nodes in the network for a given set of multicast requests. Brar et al. [24] explained the impact of interference in WMN routing. Herein, the communication network is divided into a number of subnets, and the interference within each subnet and between subnets is discussed.

In our previous work [25], the HWMP algorithm was implemented with a load-balancing strategy to avoid congestion at highly demanded smart meters. For a better understanding, the load-balancing algorithm is explained using a case study, as detailed in Section 4 herein. With an increase in the number of smart meters, packets from some smart meters require additional hops. This type of transmission is susceptible to increased delays based on the large number of retransmissions caused by packet drops. The packet drop is caused by the inability to meet the threshold value of the SINR [26] at the receiver. To reduce packet drop, a genetic algorithm is applied in this work to optimise the load-balanced paths. As a result of such optimisation, a fast and reliable transmission is possible with a reduced network load.

The paper is organised as follows: Section 2 describes the AMI architecture with the WMN factors that deploy the AMI network. Section 3 explains the problem formulation for the proposed optimisation. Section 4 discusses the LA-HWMP method with a case study. Section 5 describes the proposed GA optimisation of the LA-HWMP routing protocol. In Section 6, the realisation of the AMI network using the WMN along with GA-LA-HWMP is discussed. In Section 7, the experimental setup of the proposed method is given. The same section analyses different performance metrics under different constraints. Section 8 concludes the paper.

Section snippets

AMI communication

In this work, the AMI is deployed using a WMN where smart meters are represented by mesh routers and AMI data collectors by mesh gateways. A simple AMI architecture with a WMN is shown in Fig. 1. Mesh topology is used in the AMI network because of the correspondence in configuration, and the functionality of the mesh network with AMI [27], as listed below:

  • In the AMI network, all smart meters cannot access the gateway directly. In such cases, a smart meter can establish a path to the gateway

Problem formulation

Table 1 summarises all the symbols used in this section, along with a description of each.

The AMI communication infrastructure developed using the WMN can be represented as an undirected graph, G=(V,E), where V=MW. The set of vertices V contains the set of smart meters M and the set of data collectors W. The graph is undirected and need not be complete, as random mesh network topology is used. The task is to identify a load-balanced path pmw between a meter mM and a data collector wW with

Load-Aware HWMP (LA-HWMP)

The IEEE 802.11s uses a HWMP as the default routing protocol. The airtime metric Ca is used to select the nodes for the path, as given in Eq. (8). Ca=[O+Btr]11efwhere O is the overhead latency, Bt is the test frame size, r is the data rate in Mb/s, and ef the frame error rate. The link with the minimal airtime metric is selected to forward the packet. In our previous work [25], a load-balancing strategy was added to the HWMP and named the load-aware HWMP (LA-HWMP). In the WMN, three types of

Genetic algorithm-based optimisation

The genetic algorithm [16] is an optimisation tool that uses a population of solutions, rather than a single solution, to reach the optimal solution. The basic genetic operators are encoding, selection, crossover, and mutation. Encoding is the representation of chromosomes (decision variables) based on the application. The representation can be binary to represent the on–off state of an entity, or an integer to represent a unique value, or a string. Each individual is assigned a fitness value

Genetic algorithm for optimising la-hwmp paths

The HWMP is the recommended routing algorithm for the WMN based on the IEEE 802.11s standard [32]. By default, the HWMP uses two route identification methods, called proactive and reactive modes for static and mobile nodes, respectively. Given that smart meters are static, a proactive routing is used. This method uses the radio metric known as air time metric for path selection. All gateways initialise the route identification process by sending a route announcement (RANN) packet to all mesh

Experimental setup

The proposed GA-based algorithm was implemented using the Proto C language in OPNET modeller 14.5. The process of the GA implementation is given in Fig. 5. The network parameters used to implement the scenarios are given in Table 2.

Four scenarios with 100, 200, 300, and 500 smart meters were simulated to analyse various network parameters such as the end-to-end delay, load, and throughput.

Conclusion

In this paper, the load-balanced paths were optimised using a GA. The routing method was tested for a number of smart meters and acceptable results obtained in line with ANSI standards. The primary objective of minimising the end-to-end transmission delay was achieved by applying the GA on a set of LA-HWMP paths. This was possible by preserving links with a higher SINR value at the receiver to evolve in the next generation. This method ensured scalability, which is a key routing requirement for

References (36)

  • ...
  • YanY. et al.

    A survey on smart grid communication infrastructures: Motivations, requirements and challenges

    IEEE Commun. Surveys Tuts.

    (2013)
  • rezDiego F. Ramí et al.

    Routing in neighborhood area networks: a survey in the context of ami communications

    J. Netw. Comput. Appl.

    (2015)
  • M. Kodialam, T. Nandagopal, Characterizing achievable rates in multi-hop wireless networks: the joint routing and...
  • A. Zakaria, H. Mohamad, N. Ramli, M. Ismail, Performance evaluation of routing protocols in wireless mesh network, in:...
  • A. Quadri, K.M. Hasan, M. Farhan, E.A. Ali, A. Ahmed, Next generation communication technologies: Wireless mesh network...
  • HiertzG.R.

    IEEE 80211s: The WLAN Mesh Standard

    IEEE Wireless Commun.

    (2010)
  • X. Zhang, J. Wei, C. Qiao, Constrained multicast routing in WDM networks with sparse light splitting, in: Proceedings...
  • Cited by (11)

    View all citing articles on Scopus
    View full text