Maestro of numbers tackles wicked problems


Big data scientist, Oliver (Ollie) Chikumbo is applying powerful optimisation techniques to problems as diverse as eucalyptus management, land-use in a rural catchment and smart city planning in Portugal.

Big data and analytics are becoming crucial tools for corporations and industries such as forestry, to find answers to extremely complex questions. Big data scientists, armed with massive computing power, sophisticated algorithms and (importantly) a penchant for problem solving, stand at the forefront of efforts to create meaning from chaos.  

Ollie sits in an office surrounded by an array of continuously-running desktops, laptops and iPads. He’s the conductor of an orchestra; his computers are the musicians, his algorithms are sheets for numbers instead of notes.  

At times, a single computer might be operating for three or four days continuously before spitting out an answer to a really ‘hard’ problem.

Hard problems are the curse of mathematicians. For example, trying to optimise land use for multiple and often conflicting objectives frequently defy traditional mathematics.

Ollie’s weapon of choice, evolutionary algorithms (EA), provides the only possible solution to these sorts of problems. As their name suggests, EA borrow concepts from evolutionary genetics.  

Applying the approach to forestry, Ollie derived optimal thinning regimes for a eucalyptus forest. The problem he set out to solve was how to optimise thinning activities that maximise both sawlog value and pulpwood production volume at the same time. Typically, thinning regimes are designed to maximise either value or volume production, but not both.

Ollie’s solution would enable a forester to supply all markets and with the flexibility to provide the product most in demand at the right time. The result was a mosaic of thinning regimes across the forest estate to meet sawlog and pulpwood demands from frequently uncertain markets. These sorts of skills will be in high demand as large amounts of forestry, GIS, genotype, product, transportation and wood processing data become routinely available.

These skills have made Ollie valuable to designers of future city ecosystems such as British-based LivingPlanIT. Ollie’s techniques to handle big data can be used to optimise thousands of complex variables (for example,
balancing aesthetics with efficient use of capital), to design, build and operate cities.

Ollie envisages massive amounts of data from the future forest industry will be handled using these techniques to provide practical day-to-day  decision support for forest managers.

Sparked by a discussion with Professor Sir Peter Gluckman, and by cleverly adapting recent concepts from evolutionary biology, researchers from the Liggins Institute and Ollie have also recently teamed up to tackle ‘wicked’ problems. We thought hard problems were a challenge says Ollie – but wicked problems make them look like child’s play!