Container Stacking And Retrieval Prototype Using Genetic Algorithms

Abstract

Container stacking problem is an NP (Non-Polynomial)-Hard problem. This type of problem requires robust and adaptive algorithms, such as the Constraints Programming technique. Rule-based programming is ineffective due to the complexity of container stacking problem. The container stacking processes are restricted to certain heuristic rules, which is difficult to be coded in algorithmic ways. Improper stacking will cause non-economical container forklift movement. As a result it will increase the operational costs. This paper proposed a genetic algorithms technique to solve container-stacking problem, where a prototype has been developed. The scope of the problem is based on Kontena Nasional (KONNAS Malaysia) case study, which is restricted for one-year operational basis. The container stacking mechanism is operated under several Genetic Algorithms operators such as Selection, Mutation and Crossover. The prototype is coded using C++, ORACLE 7.0 as database and runs under UNIX platform. The average optimal stacking result obtained ranged between 78-83 percent. This prototype was developed at Artificial Intelligence System Development Laboratory (AISDEL), Sirim Berhad. Key words: Genetic algorithms; constraint programming; container stacking; heuristic rules; port management system