Gekko is historically tied to the ADAM model (Aggregate Danish Annual Model), which for many years had been solved with a really fast Fortran-based software system called PCIM. Around 2008, Gekko was developed as an alternative, and the first model to use Gekko was EMMA (Environmental eMission Models for ADAM), using Gekko from 2008 and onwards.
In 2009, Gekko was used to solve ADAM, too, and in the beginning most of the development had to do with implementing features known from PCIM. That is, most of the early focus was related to modeling. Later on, however, the focus also included data management capabilities, because the AREMOS system was being discontinued. AREMOS was (and is) being used quite a lot in Danish institutions for timeseries data management systems, so it would be practical if Gekko could implement some of these features, too.
This led to a kind of marriage of the best ideas from PCIM and AREMOS, respectively, trying to unify these in an integrated system. PCIM operates with two data banks, reflecting a baseline/reference simulation together with the current (alternative) simulation. This useful way of managing scenarios makes comparisons between scenarios much easier. AREMOS, however, does not include such concepts, but instead relies upon a list of open databanks, where the first databank on the list corresponds to the current ‘workspace’. This is practical regarding data management systems, but unifying these different capabilities took some time and effort.
Gekko 2.0 presented such an unified framework, providing an intuitive and productive environment for daily use.
The up-coming Gekko 3.0 (due in the autumn of 2018) tightens up the syntax of Gekko 2.0, and adds more flexible data structures. This will problably be the “definitive” syntax and logic of Gekko for a long time to come.
See more under ‘Features‘.