Evolutionary Multi-Objective Optimization and Preference Modeling in Green Logistics

This master thesis considers supply chain design in green logistics. The research consists of two parts. First, we formulate the environmentally conscious design as a a multi-objective optimization problem and construct the Pareto front using scalarizing methods (weighted sum and -constraint method) and genetic algorithms (NSGA-II/SPEA2). The second part involves constructing a preference model to aid the decision maker (DM) in choosing the preferred alternative using the UTAGMS method. This research includes a case study for a supply chain in the South Eastern Europe region; it extends the work of I. Mallidis, R. Dekker, and D. Vlachos. The impact of greening on supply chain design and cost: a case for a developing region. Journal of Transport Geography, 22:118–128, 2012. We apply a genetic algorithm to optimize simultaneously cost, CO2 emission and Particulate Matters (PM – also known as fine dust), and to present a set of alternatives to the DM (the Pareto front). In this case study there are two different scenarios: both have the distribution centers outsourced, one also outsources the transportation while the other scenario leases the transportation. First we compare the different method to see which method give the best representation of the Pareto front. Then the UTAGMS method will be used to aid the DM in choosing his/her most preferred solution. For the UTAGMS method, the DM is asked to provide his/her preference information by means of pair wise comparisons. Some computational tests are used to determine how applicable the UTAGMS method is to this particular case.