**4. The decision aided tool**

In this chapter the framework of the software is discussed, thus containing the architecture and the information flow. Finally, the application of the AGV (Automatic guided vehicle) in the optimization of hospital logistics is also illustrated. The tool has as functionalities to:

*How to Improve Hospital Flows in the Context of the COVID Pandemic DOI: http://dx.doi.org/10.5772/intechopen.98672*


The decision aided tool will combine different artificial intelligence concepts (expert systems, machine learning, multi-agent systems) with reasoning such as CBR, generalization, and transformation.

## **4.1 Expert system and machine learning**

To allow hospitals to optimize the internal logistic process, a decision-making tool is being developed. Combining the concepts of Industry 4.0, Logistics 4.0 adapting them in the context of healthcare. The tool being developed is based on the use of Artificial Intelligence utilizing machine learning tools and expert systems.

In this way, the decision aided system will be elaborated by having the ability to ratiocinate and learn in its own way. For this, Case-based reasoning (CBR) system will be combined with the multi-agent system, which makes it possible to achieve this specificity in the tool. With this, the system can suggest solutions to problems or give proposals for improvement, reusing or adapting the approaches developed in previous applications.

Then, machine learning, through the repetition of a behavior, will contribute to the decision-aided tool by generating new inputs and in this way progressively enriching the database.

#### **4.2 Software framework**

The software framework is divided in 3 main parts, which are, the information input, generation of the current process with the problems and possibilities of improvement, and finally the generation of the process optimized with the results of the optimization.

As the final objective of the tool is to enable an optimization of the internal logistics of the hospital, eliminating inconsistencies in the processes and reducing the non-value-added time, the inputs required by the tool will be:


With those inputs, the systems can through machine learning and expert systems, compare the data collected in hospitals with the referring hospital to achieve the ideal logistic model.

This ideal logistics model will be developed after applying Lean Manufacturing and Logistics 4.0 methodologies in several processes in different hospitals, and it will be progressively improved through machine learning.

After the analysis, the second stage of the tool will generate a series of data and the simulation relating to the process, giving the user a general idea about the logistics efficiency. The software will present the problems encountered and make suggestions for improvement to the user.

In the third and last step, after the selection of the improvements, the software will implement them in the process. For this, the system will again need to use machine learning and the expert system to make the changes and generate the data regarding the optimized process.

With this decision aided tool, the user will be able to test changes within the process and see what the results will be, facilitating the logistics optimization process within hospitals as it will reduce time and cost and increase the effectiveness.

### **4.3 Architecture of the software**

The **Figure 7** represents the architecture of the model, it aims to organize the components, modules, and information flow of the software in a visual way. Since the decision aid tool will have as functionalities, learn, and update its database in an automatic way, where several components will act, a structure was chosen where the database is centralized and combined with the tool manager.

This tool is composed of the following components:


**Figure 7.** *Architecture of the software.*

