The IBM Research Constraint Solver models and solves complex problems through the examination and prioritization of myriad variables and constraints
Variables can be a headache. Ask any seventh grader puzzling over a2=9. When variables for a given problem include the myriad manufactured parts that make a truck engine run or the skill sets of thousands of employees, the capabilities of the human mind can be overwhelmed.
Enter the IBM Research Constraint Solver. This core technology is offered by IBM Research Services, a partnership between IBM Research and IBM Global Business Services (GBS). The Constraint Solver models and solves problems involving thousands of variables and multi-tiered constraints. A robust general-purpose tool in use for more than a decade, it employs both deterministic and stochastic searches to develop high-level parameters for each problem. It then uses algorithms to solve them.
The Constraint Solver employs open interfaces, allowing it to be easily used for such diverse problems as hardware verification, workforce management, automotive configuration and circuit design. All the algorithms embedded in the generic Constraint Solver technology can be applied to each problem. Two industrial-scale applications of Constraint Solver technology now in the works involve vehicle configuration and workforce management.
Modern vehicles are highly configurable and must be shipped according to strict specifications. In addition, many types of standards, including manufacturing and engineering best practices, and international commerce laws, govern their construction and sale. In a hotly competitive market, these manufacturing complexities are coupled with tight business pressures. The state-of-the-art technology offered by the Constraint Solver can help provide valid configurations that conform to user requirements and can assist in validation and extension of orders. It also offers powerful capabilities to help solve full-fledged constraint-satisfaction dilemmas.
Human resource managers face similar challenges when deploying human capital. In a typical workforce-management problem, thousands of employees must be assigned to jobs. Variables and constraints include each individual’s skills, pay level, geographic location and availability. As companies seek to optimize their workforces, they find a growing need for tools to help forecast staff demand, place employees, address skill gaps, plan schedules and engage in long-term planning.
The Constraint Solver meets these challenges by handling situations in which, for example, human resource managers need to fill a job that requires a predetermined skill set, a set rank within the corporate hierarchy, and an employee already living in a given locale. The Constraint Solver first prioritizes constraints and variables according to client need. The tool will then either match an employee to the job, or determine whether the problem cannot be solved with a given pool of employees because of conflicting constraints.
To see how the Constraint Solver works in the real world, consider a recent GBS engagement with a major truck manufacturer. The manufacturer needed to optimize its commercial-vehicle sales chain. In doing so, it wanted to consider all aspects of engine-configuration with the goal of improving quality and manufacturing efficiency.
The IBM Research Constraint Solver proved an integral part of this GBS solution. IBM delivered a thorough upgrade of the company’s infrastructure, employing the Constraint Solver to tackle problems such as order validation and correction. Upon full implementation, the truck manufacturer will be able to better comply with governmental standards. It will be able to more quickly change configuration models as governmental standards change. Finally, the manufacturer may enjoy an improvement in its customer-satisfaction levels.
To find out more about the IBM Research Constraint Solver and to learn about how it can help solve complex problems in your business, Contact IBM Research Services today.