Leading Companies Make Use of Mathematical Optimization


Leading companies make use of mathematical optimization to improve decision-making. Mathematical optimization provides a way of solving complex problems using data and advanced algorithms. Eighty-five percent of Fortune 500 companies have been using mathematical optimization for a long time (Winston & Albright 2012). As in the case of other artificial intelligence solutions, mathematical optimization is becoming increasingly important to competitiveness as the ability of businesses to harness data improves by leaps and bounds. Many Finnish companies are also using optimization for preparing work rosters, production planning, etc.

When used for business planning, mathematical optimization also provides extensive reporting on performance metrics, cost structure and employee calendars. At the same time, it makes it easier to compare scenarios and reporting plans. As a result, people gain better insight into business and the transparency of the planning process increases.

Resource Optimization Takes Effort

Efficient use of resources is vital to the success of any business. Especially when it comes to roster planning, it is often a problem to find the right workers for the right shifts. As a result, people tend to settle for a workable solution. With a limited workforce, the number of potential combinations of duties grows exponentially – some better than others. A roster planner needs to take into account the different types of shifts, working hours regulations, extra payments, leaves, employees’ wishes and a wide range of other things.

Since roster planning is so complex, there is almost inevitably a trade-off between speed and optimum planning. The human mind is simply unable to process millions of potential combinations systematically. The roster planner’s life is a painful struggle and even so the final outcome is far from perfect in terms of efficiency in the use of the human resources.

Mathematical Optimization Helps Speed Up Planning and Improve Quality

A radical change can be achieved if a company adopts mathematical optimization. With this approach, the roster planner verifies the accuracy of the input data while an algorithm assigns people to the shifts. The process is based on the preferences specified by the roster planner as well as advanced mathematical calculations. For example, the roster planner can specify the maximum number of overtime hours to justify the rejection of a request for a day off made by an employee.

Mathematical optimization helps speed up planning and improve quality. This allows the specialist to focus on comparing the alternative plans, considering employee requests and developing operations. Moreover, he or she is able to respond to changes quickly as it only takes a fraction of the time previously required to create a new plan. For example, if an employee falls ill, a new plan can be generated by changing the availability data for that employee and clicking a button. The algorithm takes care of the reallocation of workers.

Man and Machine – A Winning Combination Of Computing Power

Of course, mathematical optimization alone is no ticket to dream land. While it is able to identify the best solution based on the input data, it does not render the planner superfluous. The roster planner is still the best expert to make the final decisions and respond to employees’ wishes as a human being. If necessary, they can also manually edit the plan created by the optimization algorithm, for example, when special arrangements are called for.

Together, the planner and algorithm form a powerful team combining a human approach, creativity, logic and tireless computing power. Optimization solutions are able to cater for a wide range of specific needs of individual businesses. Matters can be addressed down to the smallest detail and the solutions can be easily scaled up without any need for further software development efforts.

Interested in optimization? We build business planning and optimization solutions for our clients using technologies such as IBM Planning Analytics and IBM Decision Optimization.

Look at practical examples of optimization problems and solutions in “Basics of Optimization” -webinar recording.

Visa Linkiö
Lead Data Scientist