Minimizing makespan on flow shop scheduling using Campbel Dudek and Smith, particle swarm optimization, and proposed heuristic algorithm

Abstract

Scheduling is defined as the process of allocating resources over time to perform a collection of tasks within a certain period of time. PT KHI Pipe Industries in Cilegon Indonesia, has been producing high quality longitudinal and spiral weld steel pipes. The company uses FCFS (first come first serve) rule for scheduling their jobs, which means that job processing based on the first incoming prioritizes orders. In October 2018 there was delay in their production, namely 1771 pipes are still in work in process (WIP) of total demand is 6, 296 pipes. Based on these problems, this research tries to determine a better production scheduling by getting minimum makespan as scheduling performance. Heuristic and metaheuristic methods used in this research for minimizing the makespan, there are Campbell Dudek Smith (CDS) for heuristic and Particle Swarm Optimization (PSO) for metaheuristic. We also proposed a new heuristic algorithm for this problem. The results show the makespan of each method, there are CDS method: 742.87 hours, PSO method: 711.96 hours, Proposed Heuristic Algorithm method: 694.64 hours, and existing production scheduling of FCFS rules is 746.43 hours. Based on the results, the Proposed Heuristic Algorithm has smaller makespan than the others.