i-Stow : enabling stowage planning service over the web

• Main Author: Aung, Kyaw Win
• Other Authors: Hsu Wen Jing
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Computer science and engineering
• Online Access: http://hdl.handle.net/10356/59038
• Institution: Nanyang Technological University

Since Singapore is one of the major hubs for maritime logistics in Asia, it is crucial to be efficient in planning for containers arrangement, loading and unloading for different sizes and kinds of vessels. It is quite a challenge for logistic operators to produce a feasible stowage plan for huge amount of containers. Although a previous team has already finished a program called “AutoStow” for automated stowage planning, high computation power and time is needed for generating results for large number of containers arrangements. The current AutoStow system also requires an IT technician to maintain the stowage planning program on site. This project aims to solve this problem by developing a cloud service for industrial clients who need help for generating feasible plan with quick and reliable result. The cloud service is hosted in a high performance computing server. Application maintenance and upgrades are handled by the developer team and hence it will reduce the burden on industrial logistic teams. Since it is a web service, no special technical knowledge is required and one stop web portal provides easy access for everyone who needs to use “i-Stow” application which is serviced in cloud. i-Stow is developed in programming language JavaEE and hosted in Enterprise version Tomcat Server. It is implemented with web application for automated stowage planning, job request schedulers, and load balancers. Scheduler is implemented using customized shortest job first (SJF) scheme and load balancer makes use of Java RMI technology to distribute the workloads among application servers. Moreover, i-Stow application server is designed and implemented under multi-threading framework in order to achieve higher performance throughput by exploiting parallelism architecture on multi-core processers. Furthermore, the testing on web server is carried out by stimulating job arrival time by Poisson arrival distribution with different means. Performance analysis is done on these test results and illustrated in this report.