Mathematical methods support product development

Feb 28, 2005
Cimcorp Headquarters

Cimcorp develops order picking systems in co-operation with the Technical Research Centre of Finland and the Tampere University of Technology.

An old proverb The more you get, the more you want also applies in the development of logistics and order picking systems. Automation has already greatly enhanced the quality, speed and correct timing of delivery of goods from suppliers to commercial centres. However, that is not enough. The picking process has to be boosted further and the quality of the picked pallets and stacks improved.

To achieve these goals, mathematical models were developed in Cimcorp to describe the picking process. Cimcorp started co-operation with the Pori Unit of the Tampere University of Technology and the Technical Research Centre of Finland.

Robot path optimization and picking planning models have been developed during the two-year co-operation project.
When seeking a method to optimize the path of the robot, we were up against a problem that has kept the scientific community occupied for centuries: the travelling salesman problem, i.e, what is the shortest route when picking crates from different storage stacks with the gripper.

'The challenging goal was to find the shortest route. In the Travelling Salesman Problem the time to calculate the shortest route greatly increases as a function of the cities visited. For example, the number of combinations for 12 picks is 479 001 600. If we try to find the best solution just by checking all the possible routes and if we can study one route per millisecond we would spend 133 hours of calculation time. Correspondingly, the number of combinations for 18 picks is 6 • 10^15 and the calculation would take 200 000 years. In spite of more powerful methods, the greatest challenge is to keep the calculation time short', says system specialist Paavo Ranta from Cimcorp.

The picking planning model tries to solve which kind of stacks the crates are placed in to utilize the robots as efficiently as possible. The model controlling operation calculates which products are put onto each stack and which robot picks them according to robot utilization, orders and storage balances.

'We believe that we will also get other benefits from mathematical models when optimizing system operation and material flow,' explains application specialist Sakari Mikkola. 'A good example of new challenges is the shelf bay code used in shops. Products should be picked in a certain sequence so that it is easy to place them onto the shelves.'

'An intelligent model that plans the stacks and robot work load will reduce the number of mechanical parts and costs when implementing the picking system', confirms Sakari Mikkola.

This site uses cookies to provide you with a best browsing experience. By using this site you agree to our use of cookies as explained in our Privacy Policy.