Now we explain the process of computation in detail. We use
\(i_0=1\) to note the output cell. For initial configuration, the multisets of edges are
\(\{\lambda \}\) and the node cells have the following multisets.
$$\begin{aligned} \{a_1\alpha _s\alpha ^{N-1}\alpha _0^{N_1}\gamma _0^{N-1},a_2\alpha \beta ,\ldots ,\alpha _N\alpha \beta \} \end{aligned}$$
(4.12)
The the running node of the neural network is
\(i=1\). To illustrate the process more clearly, we assume that the current running node is
\(i=2\) after the first round of running for
\(i=1\). In the first node cell, the symbol
\(\pi _\sigma \) notes the number of running rounds for the network. Hence the current configuration is
$$\begin{aligned} \{a_1\alpha \beta \pi _\sigma ,a_2\alpha _s\alpha ^{N-1} \alpha _0^{N_2}\gamma _0^{N-1},a_3\alpha \beta ,\ldots ,\alpha _N\alpha \beta \} \end{aligned}$$
(4.13)
Now for each
\(1\le i\le N\) and
\(i\not =2\) we run the rules in
\(F_2\) and the node cells are
$$\begin{aligned} \{a_1\beta \pi _\sigma ,a_2\alpha _s\alpha _0^{N_2}\gamma ^{N-1},a_3\beta ,\ldots ,\alpha _N\beta \} \end{aligned}$$
(4.14)
At the same time the edge cells
\(\sigma _{2i}\) for
\(1\le i\le N\) are
$$\begin{aligned} \{a_2a_1a^{w_{21}},\lambda ,a_2a_3a^{w_{23}},\ldots ,a_2a_Na^{w_{2N}}\} \end{aligned}$$
(4.15)
Next we run the rules in
\(F_{01}\) and the edge cells remains the same while the node cells are
$$\begin{aligned} \{a_1\beta \pi _\sigma c,a_2\alpha _0^{N_2}\gamma ^{N}b_+^{w_{22}},a_3\beta ,\ldots ,\alpha _N\beta \} \end{aligned}$$
(4.16)
Next we run rules in
\(F_1\) on edges, and all the strings change to
\(b_+\) of
\(b_-\) which is noted by
\(b_{+/-}\).
$$\begin{aligned} \{b_{+/-}a^{w_{21}},\lambda ,b_{+/-}a^{w_{23}},\ldots ,b_{+/-}a^{w_{2N}}\} \end{aligned}$$
(4.17)
The next step is to run
\(F^{ij}_1\) on edges and the edge cells change into empty multisets and these states remain till the end of this round.
$$\begin{aligned} \{\lambda ,\lambda ,\lambda ,\ldots ,\lambda \} \end{aligned}$$
(4.18)
Meanwhile the node cells are
$$\begin{aligned} \begin{array}{l} \{a_1\beta \pi _\sigma c\alpha _{+/-}^{w_{12}},a_2\alpha _0^{N_2}\gamma ^{N}b_+^{w_{22}}\alpha _{+/-}^{w_{12}}\alpha _{+/-}^{w_{23}}\ldots \\ \alpha _{+/-}^{w_{2N}}, a_3\beta \alpha _{+/-}^{w_{23}},\ldots ,\alpha _N\beta \alpha _{+/-}^{w_{2N}}\} \end{array} \end{aligned}$$
(4.19)
Now we run rules
\(\{r_{01},r_{02},r_{03},r_{04},r_{00}\}\) in
\(F_0\) and count the number of
c in
\(\sigma _1\).
$$\begin{aligned} \{a_1\pi _\sigma c^{N+N_{2}},a_2\gamma ^{N}b_{+/-}, a_3,\ldots ,\alpha _N\} \end{aligned}$$
(4.20)
Now we compare the resulting
\(b_{+/-}\) with
\(a_2\) and then
$$\begin{aligned} \{a_1\pi _\sigma c^{N+N_{2}}\pi _0^x,a_2\gamma ^{N}, a_3,\ldots ,\alpha _N\} \end{aligned}$$
(4.21)
Actually in one running we only obtain one possible correct node which is noted by the symbol
\(\pi _0\). If after all the
N nodes are processed we obtain
\(\pi _0^N\) correct nodes, then the network is stable. We then send a symbol
\(\pi _h\) to the output node the computation is complete. Otherwise we send a continuing symbol
\(\pi _c\) to
\(\sigma _1\) and continue to the next step. The function of rules
\(F_3\) and
\(F_5\) is to control the running of current node to the next node.