1. Compute q-ROF score and shift to positive range: s_ij = μ_ij^q - ν_ij^q ∈ [-1,1]; s̃_ij = s_ij + 1 ∈ [0,2]. Shift ensures positive values for vector normalization.
sij=μijq−νijq∈[−1,1];s~ij=sij+1∈[0,2]
2. Vector normalization on shifted scores: s̄_ij = s̃_ij / √(Σ_i s̃_ij²).
sˉij=∑i=1ms~ij2s~ij 3. Ratio System: weighted net assessment y_i* = Σ_{j∈Ω_b} w_j s̄_ij - Σ_{j∈Ω_c} w_j s̄_ij. Rank descending.
yi∗=j∈Ωb∑wjsˉij−j∈Ωc∑wjsˉij;rank descending
4. Reference Point (optional): r_j* = max s̄_ij (j∈Ω_b) or min s̄_ij (j∈Ω_c). Chebyshev distance: y_i** = max_j w_j|r_j* - s̄_ij|. Rank ascending.
rj∗={maxisˉijminisˉijj∈Ωbj∈Ωc;yi∗∗=jmaxwj∣rj∗−sˉij∣;rank ascending