Artificial Intelligence-Based Fuzzy Logic with Modified Particle Swarm Optimization Algorithm for Internet of Things-Enabled Logistic Transportation Planning


Logistics are modified similar to the changes in retail sales. Prior to using e-commerce, the establishment of logistics was classified into three stages according to the alterations in logistics providers [1]. Industrial Revolution has resulted in several manufacturing organizations, which have been modeled in a way as to manage the record as well as transportation of the products. The economic globalization that promotes social division has led to the practice of outsourcing of logistics by producers and sellers to the third-party logistics firms having expertise in the field. This reduces the costs and improves efficiency. With e-commerce, the tasks of logistics have undergone change, serving as links between production and sales by delivering consumables directly to the users. For example, the network-based delivery organizations as well as warehouse and logistics enterprises are significant links between producers and clients [2].

Big data analytics is a technique that is applied to define a massive amount of data collection with respect to acquisition, memory, management, and analysis; this technique exceeds the abilities of existing database software tools [3]. Big data has been characterized by high-volume data, rapid information flow, variety of data, and minimum rate of density [4]. For example, number space has been employed by the databases of social network site Facebook per day; the New York Stock Exchange produces 1 ТВ of novel trade data in a single day; it generates an individual jet engine that is capable of producing 10 ТВ of data within a limited time frame. The importance of big data model can be estimated by processing the data that is used to analyze the included values instead of processing massive data. In various cases, with adequate data, a system could be developed with an application technique that guides the machines to perform smart objectives under the applications of different machine learning (ML) techniques [5].

The digitalization of logistics would be enhanced vitally where several models of labor division, like crowd sourcing and crowd funding and sharing, are applied widely. The service economy and experience economy are improved further, as AI models are being evolved in a rapid manner; thus, the “Intelligent Revolution” would modify the logistics industry. Big data is more rapid in transforming business formats and lifestyle modifications, and helps in social and economic development trajectories, which is embedded with maximum positive impact on logistics industry. With respect to the application as well as industrial developments of methodologies, existing logistics are simplified by longer multiple link chains, automatic works, and locally optimized deployments that acquire immediate modifications to face the issues caused by the modifications in a market platform for developing logistics industry [6]. In last decades, a product has been launched and used by the users, which consumes major steps at the time of selling at least five products on average. Based on the research work, to reach the customer, processing and manufacturing should be of lower duration, whereas managing and transportation should be optimum. A longer chain as well as several links of classical logistics makes complex adjustments. The issues arising at the time of massive as well as longer processes tend to minimize efficiency and maximize expenses.

In the present-day business scenario, the deployment of models like big data, AI, and robotics tend to stimulate the basic modifications in logistics intelligence. Logistics systems and its applications lead to independent route development, exploring human visual system, and some other events, which combine the modern devices with diverse links. The actual decision-making process depends upon the experience of entirely converting AI model, where the systems attain self-thinking and autonomous decision-making action [7].

The main aim of logistics is to combine modern and professional systems that lead to developing logistics route planning, which enables reaching expert level through domain information as well as the inference engine. The traditional methods are referred to be representative techniques in evolutionary models like genetic algorithms (GAs) [8], ant colony optimization (ACO), artificial bee colony (ABC), swarm optimization (SO) [9], and so on.

According to the definition provided, the purpose of logistics route planning is to effectively explore for a logistics solution that may meet every logistics operation, such as pickup and delivery needs, and reduce the overall computational cost of logistics. In recognizing the model, the initial challenge is to define superiority of a logistics solution. It has been assumed that computational cost of logistics includes courier cost based on the working time and fuel cost of overall driving distance. Therefore, if the personnel and routing costs are less, then an optimized logistics solution is attained. On the other hand, as logistics route planning issue comes under the NP-hard issue, the scalability of a logistics platform increases the processing cost. Hence, the model of managing processing efficiency as well as solution superiority is assumed to be an alternative challenge. A logistics firm has longer duration to determine the maximum-quality logistics solution. Therefore, a real-time logistics planning technique must ensure a better practical solution; however, solution may further be enhanced on the advice of modern experts, if a logistics industry has maximum duration to obtain optimal solutions.

In this chapter, a new intelligent logistic transportation planning model is presented by the hybridization of fuzzy logic with modified particle swarm optimization (HFMPSO) algorithm. The proposed model initially used Internet of Things (IoT) based barcode reader to access the details of the package. The proposed transportation planning algorithm involves three main phases: partitioning of packages, route planning using HFMPSO algorithm, and package insertion. The HFMPSO model has been tested using a set of performance measures under diverse aspects. The experimental outcome clearly verified the superior performance of the HFMPSO model over the compared methods in a significant way.

< Prev   CONTENTS   Source   Next >