1: \begin{abstract}

2: 

3: We present analytical results for the structural evolution of 

4: random networks undergoing contraction processes via 

5: generic node deletion scenarios, namely,

6: random deletion, preferential deletion

7: and propagating deletion.

8: Focusing on configuration model networks,

9: which exhibit a given degree distribution $P_0(k)$

10: and no correlations,

11: we show using a rigorous argument

12: that upon contraction the degree distributions of these

13: networks converge towards a

14: Poisson distribution.

15: To this end, we

16: use the relative entropy $S_t=S[P_t(k) || \pi(k|\langle K \rangle_t)]$ 

17: of the degree distribution $P_t(k)$ of the contracting

18: network at time $t$ with respect to the corresponding Poisson distribution

19: $\pi(k|\langle K \rangle_t)$ with the same mean degree $\langle K \rangle_t$

20: as a distance measure between $P_t(k)$ and Poisson.

21: The relative entropy 

22: is suitable as a distance measure since it

23: satisfies $S_t \ge 0$ for any degree

24: distribution $P_t(k)$, while equality is obtained only for 

25: $P_t(k) = \pi(k|\langle K \rangle_t)$.

26: We derive an equation for the time derivative $dS_t/dt$

27: during network contraction 

28: and show that the relative entropy 

29: decreases monotonically to zero during the contraction process. 

30: We thus conclude that the degree distributions of

31: contracting configuration model networks 

32: converge towards a Poisson distribution.

33: Since the contracting networks remain uncorrelated,

34: this means that their structures converge towards an

35: Erd{\H o}s-R\'enyi (ER) graph structure,

36: substantiating earlier results obtained using

37: direct integration of the master equation and computer simulations

38: [I. Tishby, O. Biham and E. Katzav, 

39: {\it Phys. Rev. E} {\bf 100}, 032314 (2019)].

40: We demonstrate the convergence for configuration model networks with 

41: degenerate degree distributions (random regular graphs), exponential

42: degree distributions and power-law degree distributions (scale-free networks).

43: 

44: \end{abstract}