Abstract

This paper presents multiple variances of selection operator used in Non-dominated Sorting Genetic Algorithm II applied to solving Bi-Objective Multi-Skill Resource Constrained Project Scheduling Problem. A hybrid Differential Evolution with Greedy Algorithm has been proven to work very well on the researched problem and so it is used to probe the multi-objective solution space. It is then determined whether a multi-objective approach can outperform single-objective approaches in finding potential Pareto Fronts. Additional modified selection operators and a clone prevention method have been introduced and experiments have shown the increase in efficiency caused by their utilization.

Highlights

  • S CHEDULING problem plays an important role in todays science and business

  • The goal of the research presented in this paper is to verify how Differential Evolution hybridized with Greedy (DEGR) and Non-dominated Sorting Genetic Algorithm II (NSGA-II) approaches explore space in the context of multi–objective optimization

  • The goal of this paper is to present a transition from a single-objective to multi-objective approach to MS–RCPSP

Read more

Summary

INTRODUCTION

S CHEDULING problem plays an important role in todays science and business. It can be met in transportation [1], production [2], project management [3], etc. Multi-Skill Resource-Constrained Project Scheduling Problem (MS–RCPSP) is NP-hard and there are no methods capable of finding an optimal solution in polynomial time [4]. DEGR algorithm is a single–objective method, and potential Pareto Front (P F ) is created by running it multiple times. Results are evaluated and compared with a set of multi-objective measures. This paper presents the transition from a single to a multi-objective approach to MS–RCPSP and introduces modified selection operators, which have proven to increase efficiency of NSGAII.

RELATED WORKS
FORMULATION OF MS–RCPSP
NON-DOMINATED SORTING GENETIC ALGORITHM II
Non-dominated Tournament Genetic Algorithm
Clone prevention
DIFFERENTIAL EVOLUTION HYBRIDIZED WITH GREEDY
EXPERIMENTS AND RESULTS
Measures
Dataset
CONCLUSIONS AND FUTURE
Results
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call