Online Programming Judge System (UOJ)

This research conducts a study to build an Online Programming Judge system with a mechanism
to generate test cases automatically using Particle Swarm Optimization (PSO) algorithm. The
system has the function to judge programming code by evaluating the output that the program
produced. Based on the problem that it is time consuming for lecturers to manually compile, run
and verify every student programs for judging. Moreover, they also need to define test cases for
different programming exercises in order to judge student‘s code. The system is built on the
purpose to assist lecturers in Universiti Teknologi PETRONAS in judging code submitted from
students and generate test cases for every programming exercise automatically. It also helps UTP
students practice and enhancing their programming skills. In this research, details of judging
process are explored. Moreover, the mechanism of test cases generation using PSO algorithm is
deeply analyzed. The study would focus on the primary structure of PSO and the proposed
fitness function to calculate fitness value for each generated test case. There are comparisons
between manual and automatic PSO test case generation results that would be conducted to
evaluate the efficiency of the proposed method. Finally, conclusion of current results and
recommendation for future development are also stated.