Integrating Knowledge Representation Techniques for Land Use Management Optimization

The effective allocation of land resources has become imperative to address the growing sustainability challenges of the modern world. In this context, optimization tools have emerged as key players in supporting decision makers through the complexities of land use management. Among them, the CoMOLA tool stands out due to its flexibility and versatility. However, an extensive investigation of recent publications using the tool uncovered a lack of standardized domain interpretation as well as technical flaws in model implementations. This paper aims to improve the development, extensibility and performance of the optimization tool by integrating knowledge representation techniques. Firstly, an ontological classification was presented as a way to address the existing language discrepancies. Its implementation would allow the preprocessing of domain knowledge within CoMOLA and improve interoperability. Secondly, an abstract model design was introduced, along with a set of ready-to-use simplified objective functions. Finally, further streamlining was pursued through the creation of utility functions to process the data extracted from the generated map individuals, and by the exploration of an adjusted version of the new design including a preprocessing layer of algorithmic data. The time and fitness performance of these two new model structures were compared with the implementation of one published research using the CoMOLA tool. The results indicate that the preprocessing layer negatively affects the execution time of the algorithm, but neither of the revised designs compromises the quality of the solutions. Additionally, they greatly reduce the programming load on users without sacrificing the original customizability. Although the accuracy of the test results could be refined, knowledge representation techniques enhance the usability and scalability of the tool, paving the way for applications beyond land use management.