Descripción del proyecto

– INTRODUCTION
Logistics in industry is one of the most important parameters of business in terms of customer satisfaction, market share and ultimately profitability. In addition to this, Covid19 has further underlined the importance and cruciality of logistics. Companies, who could respond to demands in a quick manner have been the winners of this period.
The efficiency of logistics operations are not at optimum level today. Ever changing World wide conditions resulting from natural disasters, economic parameters (oil prices, exchange rates, demand volatility) make it harder to achieve a high performance level when the work is done only by human experience.
Machine Learning (ML), introduced in 2011 as one of the most important pillars of Industry 4.0 (I4) concept, takes the advantage of information technology. Current hardware and software capabilities offered by cloud computing and storage as well as high processing speeds enable organizations to work at much higher efficiency and profitability levels in comparison conventional businesses run solely by human experience.
This project proposes a new strategy for order forecasting and optimization of loading/unloading times for Ekol Logistics, based on ML methods using Boosted Decision Trees and Deep Neural Networks.
– MAIN OBJECTIVE
The main goals of this PhD thesis are the following:
o Understand the relevant parameters in order forecasting and optimization of loading/unloading times at the company.
o Build ML algorithms to improve overall efficiency in transportation of goods.
o Increase company profitability.
o Augment customer satisfaction.
Currently Ekol Logistics establishes their routes based on personal experiences of staff at planning departments. The introduction of ML methods will represent a clear benefit thanks to fast and accurate computer decisions.
– MATERIAL AND METHODS
Phase 1. Study existing predictive system at the company
Ekol Logistics has several years of experience in logistics sector. This company has a special focus on business between Turkey, Spain and other western countries, with several customers for whom they give warehousing and intermodal transportation services between different countries. In particular there is a cross-docking facility in Barcelona and the company is expanding with plans to operate in Girona and Tarragona in the near future.
In this phase the operating mode of the company will be evaluated and understood.
Phase 2. Replicate the results achieved previously
The PhD candidate will interact with Ekol Logistics to understand how transportation related decisions are taken between Turkey and different destinations and will reproduce proposed operations in new intermodal transportation services.
Phase 3. Work on order-forecasting and optimization of loading / unloading times using artificial intelligence (AI) tools.
ML is an application of AI that provides systems with the ability to automatically learn and improve from experience without being explicitly programmed. This approach can allow businesses to benefit from superior properties of computer algorithms and hardware compared to human being, such as unlimited data storage capacity, fast-computation power and being able to work with other systems in parallel.
Logistics is one of the most suitable sectors in business to take the unique advantages ML offers, considering the vast amounts of data involved, fast-changing business demands and conditions as well as its adaptability to digitization.
In this phase of the project a simple prototype based on Boosted Decision Trees (BDTs) will be applied to order-forecasting and optimization of loading/unloading times. Phase 4. Implement a new predictive model
In this phase a new order-forecasting and optimization of loading/unloading times model for Ekol Logistics will be implemented, based on Deep Neural Networks(DNNs).
Phase 5. Test the new predictive model
The algorithms developed in this project will be used in order-forecasting and optimization of loading / unloading times. In an initial phase the decisions will be compared to previous methods but eventually the new strategy will be used at Ekol Logistics.
Phase 6. Analyse results and prepare publications



MÁS INFORMACIÓN

Si te interesa la oferta, llena el pdf con tus datos y envía ́l a doctorats.industrials.recerca@gencat.cat