Effective Techniques for Managing Large and Unstructured Graph Database

Abstract

To extract insightful data embedded into web-based information, which is crucial for newlinevarious academic and commercialized application requirements. The study thereby newlineintroduces robust computational modelling utilizing computing knowledge graph newlinemodels from collaborative web-based unstructured information. For this purpose, newlineconsidering the problem associated with network traffic a set of procedures reduces the newlinecomputational effort to a significant extent. However Big data (BD) has a huge newlineinfluence on logistics management due to the explosion of large amounts of traffic data. newlineBig data analytics systems are required to have a modular architecture, which makes newlinethem easy to develop and fix whenever an issue arises. A big data analytic architecture newlinedeals with dynamic Big Data Optimization problems known as VRP-TW where the newlineobjective is to minimize the total cost and find an optimal path. The numerical newlinetheoretical analysis shows the effectiveness of the formulated model. A framework to newlinesolve the dynamic vehicle routing problem with time windows is proposed. This newlineproblem involves determining the minimum cost routes of a homogeneous fleet of newlinevehicles to meet the demand for a set of customers within the respective time windows. newlineIn addition, new customers can be assigned to vehicles during the execution of the newlineroutes. A framework is based on two phases: a priori where the routes are obtained for newlinethe known customers using static routing and a posteriori where routes are re-optimized newlinerepeatedly during the planning horizon either continuously or periodically. The newlineframework is validated using algorithm variants based on insertion heuristic, ant colony newlineoptimization, variable neighborhood descent, and random variable neighborhood newlinedescent, which were adapted to solve a posteriori phase. The proposed algorithm is a newlinehybrid version that combines an improved version of the ant colony systems with a newlinegraph network. Computational results show that most of the algorithms are competitive newlineregarding the state of the art