Dissertation defense (January 26, 2022): Igor da Silva Morais

Student: Igor da Silva Morais

Title: Hybrid Approaches to the Two-Stage Facility Location Problem

Advisors: Pedro Henrique González Silva (advisor)  and Eduardo Bezerra da Silva (CEFET/RJ) (co-advisor)

Committee: Pedro Henrique González Silva (president),  Eduardo Bezerra da Silva (CEFET/RJ), Diego Nunes Brandão (CEFET/RJ), Vanessa de Almeida Guimarães (CEFET/RJ), Glaydston Mattos Ribeiro (COPPE/UFRJ)

Day/Time: January 26, 2022 / 08h

Room: https://teams.microsoft.com/l/meetup-join/19%3a9fe266b1abfc45d0a40451cdf514381c%40thread.tacv2/1642426974546?context=%7b%22Tid%22%3a%228eeca404-a47d-4555-a2d4-0f3619041c9c%22%2c%22Oid%22%3a%2245114d98-ef79-4a71-8ee9-16ae0daf7646%22%7d

Abstract: In the class of supply chain problems, the Two-Stage Capacitated Facility Location (TSCFL) is defined by optimal locations for installing factories and warehouses to meet the demand of customers. The problem aims to minimize operating costs: opening facilities and the flow of products from factories to customers, passing through warehouses, meeting the capacity constraints of factories and warehouses and customers’ demand. This problem can be viewed as a simplified version for application in the context of smart cities, as can cover all the three pillars, governance, energy and transportation. To solve this problem, two hybridization are proposed of Clustering Search (CS), Adaptive Large Neighborhood Search (ALNS) and Local Branching (LB) is proposed. This hybridization is a new and interesting approach which has found high quality solutions in low computational time. In order to compare and test robustness of the proposed components another hybridization is proposed a Hybrid approach of the Biased Random-key Genetic Algorithm. To show that, computational experiments were performed using benchmark instances. The results showed that the HBRKGA outperforms the current state-of-art for the TSCFL for 44 out of 50 instances and the stability of the methods is showned in a statistical analysis, in order to test differences of the method.