A Large-Scale Group Decision-Making Consensus Model considering the Experts’ Adjustment Willingness Based on the Interactive Weights’ Determination
Algorithm 1
The k-means clustering method.
Input: The value of Qm for each DM and the number of subgroups K.
Output: The clustering results G1, G2, …, GK.
Step 1. Each evaluation information matrix Qm (m = 1, …, M) is transformed into one-dimensional vector. For instance, the transformed vectors are denoted as (m = 1, …, M),
Step 2. Set or Choose randomly K vectors as the initial cluster centers {, …, }, where
Step 3. Let each DM enter to the subgroup closest to him or her. In other words, calculate the distance d(ei, ) between the DM ei and the k-th cluster center. d(ei, ) can use the Euclidean distance given by
Then, we can obtain = d(ei, ), and the DM ei therefore should enter to the subgroup .
Step 4. Recompute the cluster results by using the member information of the current subgroups. Suppose that the new cluster centers are {, …, }, where
Step 5. Set the boundary conditions of clustering process. Setting δ > 0, and the total differences before and after adjusting clustering are called TD, where
Step 6. The judgment process. If the condition TD < δ is satisfied, the clustering process is over, and then go to the next step; Otherwise, go to Step 4.