Adaptive code generation for data-intensive analytics

Modern database management systems employ sophisticated query optimization techniques that enable the generation of efficient plans for queries over very large data sets. A variety of other applications also process large data sets, but cannot leverage database-style query optimization for their code. We therefore identify an opportunity to enhance an open-source programming language compiler with database-style query optimization. Our system dynamically generates execution plans at query time, and runs those plans on chunks of data at a time. Based on feedback from earlier chunks, alternative plans might be used for later chunks. The compiler extension could be used for a variety of data-intensive applications, allowing all of them to benefit from this class of performance optimizations.

ROBUST CHANNEL ENCODING FOR MMTC NODE IN 5G (NR)

The paper is focused on robust channel encoding for Massive machine type communication (mMTC) communication in 5G (NR). The performance evaluation of channel encoding is obtained at 5G New Radio (NR) PHY. The results show that reliable bit error rate (BER) against the poor channel condition or random fluctuated channel applied. Channel encoding algorithm as a forward error correction code (FEC) is applied on packet to packet basis to improve the BER performance against inter symbol interference. The concept of adaptation of code rate is valuable to reduce the payload effect and provide optimum solution between BER and throughput. Adaptive code rate selection is based on impact of earlier transmitted packet bit using feedback indicator.

A Parallel Mix Self-Adaptive Genetic Algorithm for Solving the Dynamic Economic Dispatch Problem

The purpose of this paper is to propose a parallel genetic algorithm that has adaptive and self-adaptive characteristics at the same time for solving the Dynamic Economic Dispatch (DED) problem that is a challenging problem to solve. The algorithm selects the proper operators (using adaptive features) and probabilities (using the self-adaptive code) that produce the most fittable individuals. Regarding operations, the choice is made between four different types of crossover: simple, arithmetical, non-uniform arithmetical, and linear. Concerning mutation, we used four types of mutations (uniform, non-uniform, creep, and enhanced apso). The choice is made scholastically, which is uniform at the beginning of the algorithm, being adapted as the AG executes. The crossover and mutation probabilities are coded into the genes, transforming this part of the algorithm into self-adaptive. The multicore version was coded using OpenMP. An ANOVA test, along with a Tukey test, proved that the mixed self-adaptive algorithm works better than both: a random algorithm, which chooses operators randomly, and a combination of operators set previously in the DED optimization. Regarding the performance of the parallel approach, results have shown that a speedup of up to 3.19 can be reached with no loss in the quality of solutions.