Parent selection

Table 10.4 Parent selection from initial population for Example 10.6...

A chemical name typically has four parts in the IUPAC system of nomenclature prefix, locant, parent, and suffix. The prefix specifies the location and identity of various substituent groups in the molecule, the locant gives the location of the primary functional group, the parent selects a main part of the molecule and tells how many carbon atoms are in that part, and the suffix identifies the primary functional group. 

Name Part 2. The Parent Selecting the Main Chain or Ring... 

Upon termination of the process the algorithm selects a compound set from the working population equal to the user-supplied library size as the library proposed. The selection of the library members is performed in a manner identical to the parent selection method described previously. 

For example, one day Arnell visited the Egyptian wing at the Metropolitan Museum of Art and became enamored of mummies He and his parents selected several ways to explore the topic of mummies, which naturally led to other related topics. Eventually, the mummy unit study took on a life of its own. Arnell s activities soon included ... 

This study implements the parent selection process through the common roulette wheel selection approach outlined by Goldberg. 

From previous trials (Heinonen and Pettersson 2002), the permutation encoding approach was found to be suitable. Exponential ranking was chosen as selection method, i.e. an exponential expression is used in calculating the parental selection probabilities. Individuals with larger fitness values receive a proportionally larger probability to be selected as parents. The following expression by Blickle (1995) is used to calculate the selection probabilities ... 

Figure 1 shows the canonical GA as developed by Holland. The canonical GA encodes the problem within binary string individuals. Evolutionary pressure is applied in step 3, where the stochastic technique of roulette wheel parent selection is used to pick parents for the new population. The concept is illustrated in Figure 2, using a trivial example with a population of four individuals. Each individual is assigned a sector of a roulette wheel that is proportional to its fitness and the wheel is spun to select a parent. While selection is random and any individual has the capacity to become a parent, selection is clearly biassed towards fitter individuals. Parents are not required to be unique and, in each iteration, fit individuals may produce many offspring. 

Using roulette wheel parent selection p/2 pairs of parents are chosen from the current population to form a new population. 

Selection pressure is defined as the relative probability that the fittest individual in the population will be chosen as a parent relative to an individual of average fitness. Too high a selection pressure and a GA will rapidly converge to a suboptimal solution. While encouraging search, a low selection pressure can result in a GA taking an inordinate time to converge. In order to control selection pressure within a GA fitness values are often rescaled when applying roulette wheel parent selection. ... 

---