GA method

The optimization approach known as GA is frequently employed in complicated and massive systems to determine results near the optimal level. Consequently, GA is an excellent technique for training a neural network model for learning. A standard GA is based on a population search method influenced by the process of natural selection that relies on the concept of persistence of the healthiest⁴⁰. GA’s primary components are (a) chromosome, (b) selection process, (c) mutation process, (d) crossover, and (e) calculation and evaluation of fitness function.

We start by arbitrarily initializing a population of chromosomes, which we typically consider as potential alternatives to scheduling for any specific task. The allocation of activities to certain machines inside that chromosome allows us to obtain a fitness value (Makespan), which is acquired. After receiving the initial population, we assess each chromosome in the group according to its unique fitness value.

A smaller makespan is always desired to fine-tune the mapping. We use an allocation scheme that statistically replicates a specific chromosome and eliminates others. At the same time, we discover that improved mappings are more likely to be repeated in future generations. At the same time, the number of individuals stays constant over each age. Algorithm 1 presents the working of the GA method⁴¹.

Algorithm 1 GA algorithm