Finding the right solutions

October 2004

A new computer language being developed by academics in the School of Computer Science and Software Engineering at Monash could help industry save hundreds of hours and millions of dollars every year. Diane Squires reports.

Putting together staff rosters, production schedules or delivery routes should be among the tasks made simpler with today's sophisticated computer systems.

Money saver: Professor Mark Wallace and Professor Kim Marriott believe the computer programming language could save time and money. Photo: Melissa Di Ciero

And while software programs have been developed to meet industry needs, the highly complex programming languages currently available mean that the software is not as user-friendly as it could be for both the programmers and the end users.

Professor Kim Marriott and Dr Maria Garcia de la Banda, from Monash's School of Computer Science and Software Engineering, are leading a team of researchers at Monash, working in collaboration with Professor Peter Stuckey and his team from the University of Melbourne, to create a more advanced computer programming language, known as HAL, that can better meet the needs of users.

Monash is already known for its work in this area. In the late 1980s, researchers from Monash and Melbourne universities, with teams from Marseille and Munich, pioneered constraint programming languages, which have since been applied in industry worldwide.

However, current languages still require highly skilled expertise to build suitable programs and are difficult to use to address large-scale industrial challenges in the areas of planning, scheduling, timetabling, routing, configuration and design. Known as constrained combinatorial optimisation problems, such challenges require computer programs to conduct complex data searches to find the best solutions.

Researcher Professor Mark Wallace joined Monash this year after working on constraint programming in industry and is now part of the HAL development team.

He says companies could save millions of dollars in resources and hundreds of work hours every year if software technology could better solve these problems.

"Current programs are slow at identifying solutions or are not user friendly," he says. "The result is that industry is using programs and other mechanisms that are not designed for this kind of problem-solving. For instance, people still use spreadsheets to put together staff rosters."

One of the requirements of addressing these issues is the need to be able to write programs quickly to meet individual needs and the ability to reuse and adapt software to future needs.

"We need to be able to adapt programs to changes in the application problem, and adapt algorithms to new problems without having to reimplement similar programs time and again," he says.

"We have to find an algorithm whereby time doesn't go up exponentially with the numbers involved. For instance, if you come up with a roster for a staff base of 100 people, it shouldn't take 10 times longer to find a solution if you have a staff base of 110 people.

"And ultimately the program needs to be easy for the average worker to use. To go from software technology to business benefit is less about technology and more about meeting the needs of the end users. It's about helping companies work more optimally."

Action

Contact Professor Kim Marriott on +61 3 9905 5525 or email kim.marriott@infotech.monash.edu.au or Professor Mark Wallace on +61 3 9905 1367 or email mark.wallace@infotech.monash.edu.au. For information on constraint programming, visit www.csse.monash.edu.au/research/OandCS/.