TRANSFORMASI CHANNEL SYSTEM KE LABELLED TRANSITION SYSTEM

Parallel system must be developed with extra precision to achieve a high level of dependability to produce error-free software. Therefore we need appropriate and formal model of system. The model is necessary for verification software in model checking. A model often used in model checking is LTS (L...

Full description

Bibliographic Details
Main Authors: , SITI MUTMAINAH, , Dr. Ing. Mhd Reza M.I Pulungan, S.Si., M.Sc.
Format: Thesis
Published: [Yogyakarta] : Universitas Gadjah Mada 2012
Subjects:
ETD
Description
Summary:Parallel system must be developed with extra precision to achieve a high level of dependability to produce error-free software. Therefore we need appropriate and formal model of system. The model is necessary for verification software in model checking. A model often used in model checking is LTS (Labelled Transition System). LTS can be generated by Program Graph (PG) which is representation of programming language. However, modeling parallel system needs to describe the communication model of interactivity between a PG and another PG. So, the model can use Channel System (CS). To solve this case, it is necessary to transform CS to LTS. So, the aim of the research is developing transformation algoritm to generate LTS from a CS. Through the concept of parallelism, the algorithm is run by exploiting the so-called SOS (Structured Operational Semantic) approach. Transformation algorithm is implemented by developing a prototype tool designed with 3 essential components, input, process and output. Input is a text file that contains a model of a parallel system in form of CS. The model will be processed by using a transformation algoritm that generates LTS as output file. Functionally, the algoritm can handle model of parallel system written by CS that consists of various form of PG containing both looping and branching. Moreover, it has ability to satisfy parallel process such as interleaving, synchronous and asynchronous. In the performance testing, we can measure elapsed time and memory used to produce states. The results of the research show that elapsed time and memory tend to increase with the growth number of state, although it depends on the PG model used in the CS.