Athens University Economics and Business (AUEB)
Serial dictatorship and simultaneous allocations of scarce resources: theory and applications
Abstract
dc:descriptionThe research in this thesis uses mathematical optimization to study resource allocation problems. More specifically, problems are analyzed that involve multiple agents with strict preferences, and processes of distributing several finite resources among these agents, in order to satisfy given demands. To solve these problems, several methods were utilized, namely serial dictatorship, and simultaneous allocation (cooperative or competitive), and then applied to three real world problems. We first apply these methods in order to study a relocation problem in which countries seek to satisfy demand for reallocation from heterogenous social groups. Thus this problem consists of allocating a given number of refugees -- who are heterogeneous with respect to country of origin and characteristics such as gender, age or educational level -- from Greece to other European Union countries which have pledged to accept a certain number of refugees. In order to study this problem, a conceptual framework was developed that consists of three allocation methods: sequential multi-agent resource allocation, simultaneous allocation, and two-stage allocation. Preferences are incorporated into these methods by assuming that the destination countries have their own preferences regarding refugee characteristics, but that they also try to consider the refugees' preferences for the destination countries. While these methods vary in design and execution, all three of them aim to create a more equitable allocation methodology for both the refugees and the destination countries. These methods could also be applied to other similar types of allocation problems. Next, we model multi-agent resource allocations consisting of many agents and sites and analyze the serial dictatorship and simultaneous allocation approaches. These approaches are then applied to an actual case of mussel harvesting. For tractability, the model is developed in a multi-fisher, multi-site fishery where one fisher may emerge as a Stackelberg leader while the rest are followers; however, the generalized model can be used to study many other resource allocation problems. Network externalities are introduced as congestion costs, and management systems based on individual and collective quotas are analyzed. The presence of a Stackelberg leader is a novel approach that extends results obtained in resource allocation, indicating the possibility of full elimination of followers, with impacts on the sustainability of fishing communities. Allocation outcomes under different assumptions are compared, and policies that could prevent the followers from exiting the market are discussed. Finally, we examine and compare mechanisms for finding Pareto optimal solutions for a many-to-many allocation problem. Explicitly introducing preference ordering and providing linear programming representations of the allocation problem enables us to study serial dictatorship and competitive matching with simultaneous choices -- which is basically a generalized Nash equilibrium problem -- and to show that under certain conditions this Nash equilibrium allocation is also Pareto optimal. The paper examines the ability of our mechanisms to provide Pareto optimal allocations, and compares the solutions obtained from the different mechanisms. The allocation methods are then applied to a many-sources, many-uses water allocation problem, and the benefits associated with serial dictatorship or competitive allocations are compared. The theoretical research into the areas described above was combined in each case with an application, in order to demonstrate its usefulness in addressing real world problems and in helping to design efficient policy regarding resource allocation. Furthermore, the outcomes of the research suggested many promising areas for further research.
Degree
thesis:*- Grantor dc:publisher
- Athens University Economics and Business (AUEB)
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Xepapadeas, Petros
- Ξεπαπαδέας, Πέτρος
Subjects
dc:subject × 18- Σειριακή δικτατορία
- Βελτιστοποίηση Pareto
- Προτιμήσεις
- Ανάλυση αποφάσεων
- Nash ισορροπία
- Συστήματα πολλαπλών πρακτόρων
- Serial dictatorship
- Pareto optimality
- Preferences
- Decision analysis
- Nash equilibrium
- Multi agent based systems
- Οικονομικά και Επιχειρήσεις
- Κοινωνικές Επιστήμες
- Διοικητική Επιστήμη και Επιχειρησιακή Έρευνα
- Economics and Business
- Social Sciences
- Management Science and Operations Research
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/54212