Fig 1.
Graph representation of example 16-node synthetic networks: a) BA, b) ER, c) WS; and degree distribution charts of example 2000-node synthetic networks: d) BA, e) ER, f) WS.
Fig 2.
Visual representation of the six preference functions used in the PROMETHEE methods: a) usual criterion, b) U-shape criterion, c) V-shape criterion, d) level criterion, e) V-shape with indifference criterion, f) Gaussian criterion.
Table 1.
Formulae for the six preference functions used in the PROMETHEE methods [41].
Fig 3.
Example GAIA plane representing a viral marketing campaign strategy selection with four possible strategies and three criteria: c1—Maximization of coverage, c2—Minimization of the number of iterations, c3—Minimization of the seeding fraction.
Fig 4.
Proposed framework based on five stages: Analysis of real network (I), synthetic network selection process (II), simulations within synthetic network (III), campaign planing (IV) and multi-criteria campaign evaluation(V).
Table 2.
Mapping viral campaign characteristics into simulation model parameters and outputs.
Fig 5.
Degree distribution chart of A) the real network, B) the selected synthetic network.
Fig 6.
Visual representation of the 150 synthetic networks used to approximate the [77] real network.
Table 3.
PROMETHEE II parameters for the synthetic network.
Fig 7.
Visual representation of the 10 best strategies in PROMETHEE II rankings for the synthetic (a-c) and real (d-f) networks, based on V-shape preference function with no indifference threshold (a,d), V-shape preference function with indifference threshold (b,e) and Gaussian preference function (c-f).
Table 4.
Results of the PROMETHEE II method analysis on the synthetic network: a) V-shape preference, b) V-shape preference with indifference threshold, c) Gaussian preference.
Fig 8.
Synthetic network’s GAIA analysis of individual criteria with visible (a) and hidden (b) strategies. GAIA analysis of grouped criteria (c).
Table 5.
Stability intervals for criteria groups in the PROMETHEE II ranking with V-shape preference function with no indifference threshold for the synthetic network.
Table 6.
Stability intervals for individual criteria in the PROMETHEE II ranking with V-shape preference function with no indifference threshold for the synthetic network.
Table 7.
Stability intervals for individual criteria in the PROMETHEE II ranking with Gaussian preference function.
Table 8.
PROMETHEE II parameters for the real network.
Table 9.
Results of the PROMETHEE II method analysis on the real network: a) V-shape preference, b) V-shape preference with indifference threshold, c) Gaussian preference.
Fig 9.
Comparison of the ranks of strategies obtained based on a synthetic and a real network [77].
Fig 10.
Real network’s GAIA analysis of individual criteria with visible (a) and hidden (b) strategies. GAIA analysis of grouped criteria (c).
Table 10.
Stability intervals for criteria groups in the PROMETHEE II ranking with V-shape preference function with no indifference threshold for the real network.
Table 11.
Stability intervals for individual criteria in the PROMETHEE II ranking with V-shape preference function with no indifference threshold for the real network.
Table 12.
Stability intervals for individual criteria in the PROMETHEE II ranking with Gaussian preference function for the real network.