2.1 Biopharmaceutical cloud platform
It is difficult to find new chemical entities in biological pharmacy. This issue extended the research and development time of new drugs and increased the research and development (R & D) costs. At the same time, the interference of external environmental factors and the variability of the biological pharmaceutical platform cannot be predicted. Therefore, compared with the traditional biological pharmaceutical technology, the recent pharmaceutical research and development of the pharmaceutical industry tends to be networked and intelligent. The network of biopharmaceutical [17] was reflected in the revolutionary changes and the network management way of biological research and pharmaceutical research and development. Intelligent [18] includes pharmaceutical platform organization structure, production process, and value chain-driven combination and other aspects of the intelligent level.
In order to give full play to the network and intelligence of the driving force of biopharmaceutical, the biopharmaceutical platform is gradually changing. On the one hand, in order to reduce R & D costs, shorten the development cycle, and decentralize R & D risk, the independent laboratory model gradually transferred to the direction of cooperation in R & D [19]. On the other hand, the application of a large number of new technologies and methods would provide a comprehensive [20] experimental platform. This platform can effectively shorten the time of drug development and improve the range of drug types.
The cloud platform has the advantages of the above two aspects. The combination of biopharmaceutical and cloud platform can not only build the cloud scale of biopharmaceutical research and development platform but also provide a comprehensive interactive platform for drug experiment.
Biopharmaceutical cloud platform is a very complex cloud environment. The cloud environment is involved in different aspects of biological pharmaceuticals, as shown in Fig. 1. Specific contents are as follows:
-
(1)
Biological pharmaceutical technology with high difficulty and complexity.
-
(2)
Experimental platform of biopharmaceutical technology
-
(3)
The intelligent management of biological pharmaceutical cloud platform
-
(4)
Cloud platform for the virtual environment of various diseases of experimental drug performance, such as cancer, genetic disease, cardiovascular disease, infection, and immunity.
-
(5)
Network cloud platform for all kinds of drugs and pharmaceutical process. The adaptability of the biological pharmaceutical cloud platform can be obtained by Eq. (1).
$$ \left\{\begin{array}{l}{E}_{\mathrm{N}}=\sqrt{B\cdot {\displaystyle \sum_{i=1}^n{C}_i\cdot {I}_C}}\\ {}{E}_{\mathrm{M}}={V}^2\cdot {\displaystyle \sum_{j=1}^m{B}_i}\end{array}\right. $$
(1)
Here, E
N said network cloud adaptation. E
M denotes smart cloud adaptation. B represents the fitness of biopharmaceutical technology. C represents the fitness of the drug experimental platform. Parameter n represents the number of test objects on the platform. I
C indicates intelligent cloud weight. V represents the balance of cloud platforms. Parameter m says share of biotech drugs.
2.2 Drug green crowd architecture
There are P servers in the biopharmaceutical cloud platform. Biopharmaceutical networking requires f virtual clouds with high frequency and large memory. The cloud platform centric server can support L data link with parallel transmission. The intelligent model of cloud platform center can satisfy the K biopharmaceutical process in parallel. All biopharmaceutical industry processes are in the same network environment and cloud environment.
Drug performance vector PB = (PB0, PB1, …, PB
n-1). We need to design n virtual test cloud environments that must satisfy the requirements of drug performance testing. The real-time available memory resources and processor utilization of the cloud platform server can be obtained by Eq. (2). From Eq. (2), the more the parallel data link, the lower the utilization rate is. It is difficult to improve the structure of large-scale biopharmaceutical industry chain.
$$ \left\{\begin{array}{l}{I}_{\mathrm{M}}=\frac{\delta}{{\displaystyle \sum_{i=1}^L\varepsilon \sqrt{D_i}}}\\ {}{R}_{\mathrm{CPU}}=\frac{1}{{\displaystyle \sum_{i=1}^K\left(1-{U}_i\right)}}\end{array}\right. $$
(2)
Here, I
M is used to show the memory utilization. D represents the data size of each data link. δ denotes data link data utilization. ε represents the mutual interference of parallel data chains. R
CPU indicates processor utilization. U represents the efficiency of the use of each test server.
So, we proposed the green crowd architecture based on the green drive of drug production, as shown in Fig. 2. Here, the medicine green crowd module has five substeps, which are S1, S2, S3, S4, and S5. The data of pharmaceuticals was sent to S1. The session of pharmaceuticals was sent to S2 and forwarded to S4. The evaluation of pharmaceuticals was sent to S3 and forwarded to S5. The final results of medicine green crowd module were sent to the crowd cloud.
The architecture core in Fig. 2 is divided into five stages. According to Eq. (3), stages S1, S2, and S4 complete the green screening and reorganization. According to Eq. (4), stages S3 and S5 complete the crowd data processing.
$$ {D}_{\mathrm{g}}=\left\{{D}_1,{D}_2,\cdots, {D}_L\right\}\frac{\varepsilon {\displaystyle \sum_{i=1}^L\varDelta {D}_i}}{{\displaystyle \sum_{j=1}^K\sqrt{I_{\mathrm{M}}^2+{R}_{\mathrm{CPU}}^2}}} $$
(3)
Here, D
g represents the data sequence of the reorganization. ΔD is the mean value of data sequence after green filter.
$$ {D}_{\mathrm{C}}=\frac{D_{\mathrm{g}}}{n}\cdot \frac{1-{\displaystyle {\int}_{j=1}^L{E}_j}}{{\left\{{\displaystyle \sum_{a\to \left\Vert L\right\Vert }{P}_E^a}\right\}}_{\max }} $$
(4)
Here, D
C expressed crowd data sequence. Parameter E expresses crowd computation overhead. Parameter a represents the length of the vector for the performance evaluation of the drug. Parameter P represents a vector of drug performance evaluation.