Riassunto analitico
After the liberation of the natural gas market, the distribution services in Italy face significant competition and currently have different optimization opportunities. Research has shown that the attended home services quality provided by IRETI S.p.A is influenced by the arrival time at the site of an emergency call, which is set from the authorities and presented in a form of tender. This thesis aims to determine a model to ensure the arrival of technicians in compliance with the quality standard, to guarantee optimal allocation of resources and to measure the performance of the simulated service.
Based on a review of the literature on supervised machine learning and attended home delivery services, various tools were implemented and developed to meet the diverse requirements and problems present in the province of Parma. Municipalities were divided into two groups (known and unknown demand) based on historical data provided from the company and used to predict the total demand from each municipality in the province of Parma. Based on the knowledge generated by the prediction tool, optimization and simulation of the organizational model were carried out. The various tools are based on Artificial Neural Network algorithm and mathematical models of Integer Linear Programming (ILP).
The new organizational model with depot positioning results was simulated to test the model's credibility and the routing costs, which showed significant results.
The results indicate that the model guarantees an excellent service quality that IRETI can use to present in the tender. On this basis, the use of the methods presented is recommended. The solution also serves expansion purposes and aim to evaluate participation in tenders in currently unserved territories.
|
