The Human Resource Allocation Problem is an interesting and important problem that occurs in several walks of life. The problem involves multi-objective optimization in a huge search space and hence genetic algorithms prove to be effective in solving such problems. Among the several types of genetic algorithms, multi-population genetic algorithms have been found to be more efficient in terms of finding the optimal solution in a limited time due to its nature of independent evolution. This paper presents a novel multi-population genetic algorithm for solving the human resource allocation problem.
Keywords: Genetic Algorithm, Multi-population GA, Multi-objective optimization, Resource Allocation
Resource Allocation Problem (RAP) is one of the classical problems in OR. Resources that can be machines, manpower or anything, which are limited, have to be allocated to different number of projects, jobs, etc. RAP is a multi-objective optimization problem that usually has multiple objectives and aims at minimizing the cost and maximizing the profit or efficiency gained, simultaneously.
[...] Calculate the total cost and total efficiency for each individual i by summing up the cost which is denoted as costi and effi respectively. b. The difference between effi and costi has to be calculated which gives the fitness of individual i.e. f(i). At each generation, the individual with maximum will be the fittest individual of that generation and it is said to dominate all other individuals. Also there can be more than one such individual in which all these individuals are said to dominate other individuals. [...]
[...] Mutation: Mutation is a genetic operator which involves only one individual in the creation of a better offspring with improved parental traits and is identical to the parent. Mutation ensures a more complete coverage of the search space by randomly altering values on genomes. Here we used swap mutation which randomly selects two positions in the chromosome and swaps the values in those positions. The working of swap mutation is explained in Figure 3. Before Mutation After Mutation Figure Working of Swap Mutation vii. [...]
[...] Migration enables even low fit individuals to mate and produce better offsprings thus ensuring a large search space to be explored Proposed Multipopulation GA In this paper, we present a multi-population genetic algorithm for solving the multiobjective human resource allocation problem. The pseudocode for the proposed MGA is as follows: 1. Generate initial population randomly Evaluate the fitness of each individual based on the non-dominance with other individuals Sort the population based on the fitness of the individuals Divide the population into desired number of groups For each group do, a. [...]
[...] Chi-Ming Lin and Mitsuo Gen proposed a hybrid genetic algorithm for solving the human resource allocation problem based on the multistage decision making model. Genetic algorithm approach which is proved to give better solutions for optimization problems is adopted in this paper. Genetic algorithm is an adaptation procedure based on the mechanics of natural genetics and natural selection, Genetic algorithms operate with a population of possible solutions rather than a single candidate and are able to evolve and give better solutions. [...]
[...] This ensures that there is no better individual in the search space Results and Statistics The Human Resource Allocation Problem has been solved by implementing the proposed multi-population GA and better results are obtained when compared to results given in The algorithm is also tested with few more test cases and found to give good results. The inputs taken from are given in Tables 1 and 2. Table Expected cost cij Number of Jobs: i Workers: j Table Expected efficiency eij Number of Jobs: i Workers: j Table Comparative results moHGA Overall jx4j cost Proposed MGA Overall jx3j jx4j cost Solution k jx1j jx2j jx3j Overall efficiency jx1j jx2j Overall efficiency From Table we can see that the difference between the total efficiency and total cost is minimal when compared to that of the multiobjective hybrid GA proposed in Thus it is clear that the proposed GA performs well with respect to both the objectives Conclusion The algorithm was tested for the resource allocation problem with two objectives and the results proved better than standard genetic algorithm and other hybrid genetic algorithm. [...]
using our reader.