README.md

FCM-based Maturity Model for Smart Manufacturing

Implementation of a FCM-based Maturity Model for Smart Manufacturing focusing on IT systems and key enabling technologies as drivers. The map assesses (running the FCM) the level of smart manufacturing of a company to capture the current state of digitization (As Is) and to suggest (via Genetic Algorithm) paths for digital growth (To Be).

Structure of the repository

.
├── evaluation                      # sources for the evaluation
|   ├── eval_structure.py           # evaluation of the structure
|   ├── eval_fcm.py                 # evaluation of the FCM inference
|   ├── notebook_eval_fcm.ipynb     # evaluation results of the FCM inference
|   └── ...
├── model                           # sources of the FCM model
|   └── ...
├── cases                           # sources of the cases to be evaluated
|   └──...
├── utils
|   ├── data                        # raw data to construct the model
|   |   └──...
|   └── single_FCM.py               # script to find the lambda of each FCM
├── config.json                     # configuration file
├── FCM_class.py                    # FCM class implementing inference
├── FLT_class.py                    # FLT class implementing the membership functions
└── ...

Getting Started

• Install Miniconda or Anaconda if you haven't already.

• Create a new conda environment:

  conda create -n pyfcm python=3.10
  conda activate pyfcm

• Install the dependencies:

  pip install -r requirements.py

Run the code

• Activate the conda environment:

  conda activate pyfcm

• Define the activation levels (AL) of each technology (node) inside the cases folder - follow below instructions. Or use one of the cases already available [low, medium, high, mix].

• Modify config.json as you wish. As an example:

  {
      "case": "low",
      "target_val": "VH",
      "to_remove": []
  }

  case key refers to the company case to study (one of the folder in cases), target_val refers to the target value to reach in the genetic algorithm (accepted values are [VL, L, M, H, VH] = [very low, low, medium, high, very high]) and to_remove refers to the set of IT systems to remove from the map (accepted values are ['CAD, CAM, PLM', 'CRM', 'ERP, SCM', 'WMS, TMS', 'MES']).

• To run the FCM (inference analysis):

  python FCM_class.py

Define AL for a new case

• Create a new folder in cases.
• Create five files X_al.csv for X=[1,5]={1 = "CAD,CAM,PLM", 2="CRM", 3="ERP,SCM", 4="WMS,TMS", 5="MES"} to define the AL of each technology of each IT system. Linguistic terms used are [NA, VL, L, M, H, VH] = [neutral, very low, low, medium, high, very high]. The concept linked to the main FCM, the one in $(row=0,column=0)$ must have NA value. Below an example:

  NA,0
  L,0
  M,0
  H,0
  L,0
  L,0
  M,0

  To check the technologies linked to each node, have a look at the .json files in model.
• Put the files in the folder you created.

How to run the experiments

Graph theory analyses

In order to show results of the graph theory analyses:

cd evaluation
python eval_structure.py

FCM inference analysis

In order to run and plot results of the FCM inference analyses:

cd evaluation
python eval_fcm.py

N.B. Results are also reported in this notebook.

GA analysis

In order to run and plot results of the GA analysis:

cd evaluation
python eval_ga.py

N.B. This pickle file stores results shown in the article and this notebook reports the outcome. The folder ga_results contains the values found by the GA.

Case Study Results

Final activation level of each IT system:

Algorithm: Papageorgiou, Iterations: 100, Company Type: case_study
CAD, CAM, PLM (FCM1): 0.80205
CRM (FCM2): 0.74365
ERP, SCM (FCM3): 0.7686
WMS, TMS (FCM4): 0.2622
MES (FCM5): 0.72659
Smart Manufacturing maturity level: 0.660618

Experts responses

Responses of the experts are here available: link.

Final causal relationships

Causal relationships resulting from the questionnaires are here available: link.

License

Distributed under the MIT License. See LICENSE for more information.