Reasoning Engine Implementation Knowledge Base System of Criteria and Indicators

Figure 4.13. The relation between nodes

4.4.2. Reasoning Engine

The inference engine of the “hierarchical network” reasons from the leaves to the root as shown in Figure 4.14. Scores of subordinate nodes determine the score of their super ordinate. The score of a node is determined by multiplying the score of its subordinates with their relative importance. For instance, if a node has three subordinates with scores 5, 6 and 8 and relative importance 0.4, 0.4 and 0.2 subsequently, then the score of that node is 5 x 0.4 + 6 x 0.4 + 8 x 0.2 = 2 + 2.4 + 1.6 = 6 A score might represent one’s guess, containing words such as true, highly true, likely, unlikely, impossible etc. Those degrees of belief can be expressed by a real number in some interval – for example between zero and one. Such a number is known as a certainty factor. KBS combines the certainties of proposition with rules. If cN1 and cN2 denote certainties for N1 and N2, then combinations of N1 and N2 follow cN1 AND N2 = mincN1, cN2 cN1 OR N2 = maxcN1, cN2 OR relation AND relation N1 N.1.1 N.1.2 N.1.3 N.1.4 N.1.5 Figure 4.14. Assessment process

4.4.3. Implementation

The KBS has two major capabilities: adapting local knowledge and assessing forest sustainability, based on the combination of scientific and local knowledge. The formal representation of the KBS is as follows top_goalSFM is_node“SFM” includesSFM,Node 1 includesNode 1 ,Node 2 includesNode 2 ,Node 3 includesNode n-1 ,Node n attributesNode,Text_explanation,[Arguments],[Creators],[Context],Remar ks oritemsNode,[Nodes] identical_itemsNode,[Nodes] userCreator_code,[Creator_attributes],[Date],[Time] contextContext_code,[Context_attributes] basketNode_type,Node Node_typeŁ Goal|Criterion|Verifier|Sub_verifier|… assessNode,Score,Certainty_factor,Relative_importance,Remarks Symbols in the bracket, for instance [Arguments] is a list of data structure. [Arguments] is a list of arguments, which has no length limitation. Scoring process The goal of the whole assessment process is sustainability of forest or SFM. It is the root of knowledge hierarchy , represented by top_goalSFM In the next level of the hierarchy are the criteria of SFM, which are represented as includes statements, as follows: includesSFM,Normal timber products and services includesSFM,Normal non-timber products and services includesSFM,Biodiversity includesSFM,Protected areas includesSFM,Stakeholders rights includesSFM,Stakeholders learning capacity The biodiversity third statement comprises indicators, which explain how to measure biodiversity. They are represented as follows: includesBiodiversity,Landscape pattern includesBiodiversity,Richness index includesBiodiversity,Genetic diversity This hierarchy process can be followed to their verifiers or sub-verifiers whenever possible. Each item or node has attributes explaining the argumentation - who created them, remarks, context etc, as represented in the formal representation. The adaptation processes are simply whether the scientific knowledge meets biophysical and socio-economic conditions of a specific site. If the scientific knowledge cannot meet the local condition then the adaptation process is performed. So, it depends on what the user wants. The user can also use the local knowledge for an assessment of sustainability. To ensure local knowledge is sufficient, the user can use local knowledge by modifying scientific knowledge through a process called an adaptation of scientific knowledge. The adaptation processes are supported by instructions of add new node, delete the existing node, reword the node and restore the deleted node. In the adaptation Although the knowledge representation forms a network, the term hierarchy is still used to describe the leveling of the knowledge. process, the user is asked to submit an argument to ensure other users know the reason behind the adaptation process. The KBS was programmed with a computer language called PROLOG Programming in Logic, which is a well-known language for Artificial Intelligence. An environment, namely VISUAL PROLOG created by Prolog Development Center, Denmark, was used. Appendix 3 illustrates the screen show of the implementation of KBS. The ultimate goal of KBS creation was combining and structuring knowledge. It was not aimed to distribute it widely, but to create a tool to facilitate the combination of stakeholders’ knowledge on CI.

4.5. Artificial Society of Forest Actors