Search

In the distributed database systems, the relations needed by a query can be kept in several locations. This process significantly increases potential corresponding Query Execution Plans (QEP's) for a user query. Henceforth, in addition to the expense of local computing, the charge of transferring data between different cloud sites should also be considered. It does not sound logical to investigate all potential query plans in a high setting like this. The best query plan (regarding cost) must be generated for processing a given query. A new hybrid multi-objective genetic and bat algorithm, a Multi-Objective Genetic Algorithm with BAT (MOGABAT), is used in the present article to produce the best query plans. The functionality comparison is made on different join graph structures, among MOGABAT, Multi-Objective BAT (MOBAT), and Non-dominated Sorting Genetic Algorithm II (NSGA-II). The obtained results have shown that the quality of generated query plans is enhanced for the join graph structures. Nevertheless, more execution time is needed. [ABSTRACT FROM AUTHOR]

Internet of things has become an essential principle of human life with the widespread acceptance of intelligent environments, where everyday objects can communicate through the internet. Mobility is the most critical factor in today's internet of things devices to apply in real-world applications. Also, proper routing protocol plays a vital role in communication and reduces devices' energy consumption. The clustering approach is one of the efficient routing techniques to improve energy consumption and enhance network life. Due to the NP-Hard nature of clustering, a krill herd optimization algorithm is proposed in this paper to select the cluster head nodes and intermediate nodes required for routing. The simulation results using NS-3 confirmed that the proposed technique performs better than particle swarm optimization and cuckoo search in terms of network lifetime. The proposed technique increases the total network lifetime by at least 11.1% compared to the current clustering methods. [ABSTRACT FROM AUTHOR]