"Adventures in Optimization" - Part 1

My 44+ year journey in the Canadian oil & gas industry has provided many, many experiences and learnings, especially in the area of price risk management and revenue optimization for junior/intermediate oil & gas producers. In the early years, discussions and decisions around how to optimize future revenue were based on opinions and subjective knowledge, often referred to "gut feel". In a cyclical industry, where uncertainty of future prices is always a threat to company success and/or survival, forecasting prices and optimizing revenue was more of an art form.

It wasn't until I was introduced to the "science" of modelling prices that the subjective approach to consulting to producer clients transitioned to an educated opinion backed by Monte Carlo-based statistical analysis. My tools of choice were Palisade's (now known as Lumivero) @RISK and RISKOptimizer software, which are capable of handling complex revenue optimization problems and providing clients with easy-to-understand visual outcomes. I found the probability-based results facilitated better decisions compared to simple scenario analysis, which tended to be construed as each scenario having an equal chances of occurrence.

Example of results of an oil & gas revenue optimization model using RISKOptimizer software.

That said, I would be remiss in leaving the impression that this approach was always successful or right. It was and still is (in my opinion) a reasonably logical approach to use for relatively short-term optimization problems (eg. up to a year out into the future), as long as there were no fundamental changes to the market (eg. changes in export gas pipeline capacity) during the forecast period.

Example of short-term gas price forecasting model using @RISK software.

The application of this approach and software tools is not limited to revenue maximization problems. It is equally applicable to cost minimization from the perspective of an industrial operation with multiple alternatives for feedstock with material levels of price uncertainty.

Looking forward, the application of artificial intelligence and agents to problems of revenue optimization or cost minimization will no doubt provide decision makers with even more insights with which to use to manage their businesses more effectively.