OR Seminar

Date: 13/12/2017

Hour: 14h30

Venue: Sala 1.5 - edifício VII, FCT NOVA

Lexicographic solutions for a tri-objective multi-period model to redesign a food bank supply chain network

Margarida V. Pato (ISEG, CMAFCIO)

We address the problem of redesigning a multi-echelon supply chain network 

for the collection of donated products by food banks and their distribution to non-profit organizations that provide food assistance to the needy population. Important strategic decisions comprise opening new food bank warehouses and selecting their storage and transport capacities from a set of discrete sizes over a multi-period planning horizon. Facility decisions also affect existing food banks that may be closed or have their capacity expanded. Logistics decisions involve the number of organizations to be supplied, their allocation to operating food banks, and the flow of multiple food products throughout the network. Decisions must be made taking into account that food donations are insufficient and a limited investment budget is available. Sustainability is accounted for within the decision-making process by integrating economic, environmental, and social goals. To this end, three objectives are defined: (i) minimization of the cost of operating the food bank network, (ii) minimization of the environmental impacts of food waste and CO2 emissions, and (iii) maximization of the social benefits generated for organizations supported by food banks. A novel tri-objective mixed integer linear programming model is developed and a computational study is conducted on the basis of a set of test instances capturing various characteristics of the southern region of the network coordinated by the Portuguese Federation of Food Banks. Using a general-purpose solver, we identify and characterize the lexicographic solutions of the tri-objective problem. The trade-offs achieved by considering the three conflicting objective functions are also investigated. We include some notes on the work in progress.

Arc Routing Today – Theory vs Practice

Maria Cândida Mourão (ISEG, CMAFCIO)

Arc Routing is among the essential problems faced by several corporations or institutions regarding either distribution or collecting activities. Depending on costumer localization and on the characteristics of service, a NRP (Node Routing Problem) or an ARP (Arc Routing Problem) should be considered. Refuse household collection, parking meters collection or mail distribution are illustrative examples of ARP applications. 

Usually, real base problems demand for specific and new constraints. Such is the case of household refuse collection, where the working time of crews must be taken into account, different routes must be balanced in terms of the work load, and made visually attractive. When dealing with money collection safety is an issue, and routes dissimilarity must be enforced. 

These problems are presented and discussed based on real case-studies.