Optimizing Hadoop Performance for Big Data Analytics in Smart Grid

Algorithm 1

GEP implementation.

Input: A set of Hadoop job running samples;

Output: A correlation of the Hadoop parameters;

  FOR    TO size of population DO

  create chromosome with the combination of mathematic function and parameter;

   fitness value = 0;

   ++;

  ENDFOR

  best chromosome = chromosome ;

  best fitness value = 0;

  WHILE   < termination generation number DO

   FOR    TO size of population DO

    Translate chromosome into expression tree ;

    FOR    TO the number of training samples DO

     evaluate the estimated execution time for case

13     IF ABS (timeDiff) < bias window THEN

      fitness value ++;

     ENDIF

     ++;

    ENDFOR

    IF fitness value = the number of training samples THEN

       best chromosome = Chromosome   GO TO;

    ELSE IF fitness value > best fitness value THEN

       best chromosome = Chromosome ;

       best fitness value = fitness value ;

3     ENDIF

    Apply replication, selection and genetic modification on chromosome proportionally;

    Use the modified chromosome to overwrite the original one;

    ++;

  ENDFOR

  ++;

 ENDWHILE

 Return best chromosome