Detection of Optimal Models in Parameter Space with Support Vector Machines

The paper proposes an approach aimed at detecting optimal model parameter combinations to achieve the most representative description of uncertainty in the model performance. A classification problem is posed to find the regions of good fitting models according to the values of a cost function. Support Vector Machine (SVM) classification in the parameter space is applied to decide if a forward model simulation is to be computed for a particular generated model. SVM is particularly designed to tackle classification problems in high-dimensional space in a non-parametric and non-linear way. SVM decision boundaries determine the regions that are subject to the largest uncertainty in the cost function classification, and, therefore, provide guidelines for further iterative exploration of the model space. The proposed approach is illustrated by a synthetic example of fluid flow through porous media, which features highly variable response due to the parameter values' combination.

Optimal Contracts, Adverse Selection, and Social Preferences: An Experiment

It has long been standard in agency theory to search for incentive-compatible mechanisms on the assumption that people care only about their own material wealth. However, this assumption is clearly refuted by numerous experiments, and we feel that it may be useful to consider nonpecuniary utility in mechanism design and contract theory. Accordingly, we devise an experiment to explore optimal contracts in an adverse-selection context. A principal proposes one of three contract menus, each of which offers a choice of two incentive-compatible contracts, to two agents whose types are unknown to the principal. The agents know the set of possible menus, and choose to either accept one of the two contracts offered in the proposed menu or to reject the menu altogether; a rejection by either agent leads to lower (and equal) reservation payoffs for all parties. While all three possible menus favor the principal, they do so to varying degrees. We observe numerous rejections of the more lopsided menus, and approach an equilibrium where one of the more equitable contract menus (which one depends on the reservation payoffs) is proposed and agents accept a contract, selecting actions according to their types. Behavior is largely consistent with all recent models of social preferences, strongly suggesting there is value in considering nonpecuniary utility in agency theory.

Optimal Design of Permit Markets with an Ex Ante Pollution Target, August 2006

In this paper, we examine the design of permit trading programs when the objective is to minimize the cost of achieving an ex ante pollution target, that is, one that is defined in expectation rather than an ex post deterministic value. We consider two potential sources of uncertainty, the presence of either of which can make our model appropriate: incomplete information on abatement costs and uncertain delivery coefficients. In such a setting, we find three distinct features that depart from the well-established results on permit trading: (1) the regulator’s information on firms’ abatement costs can matter; (2) the optimal permit cap is not necessarily equal to the ex ante pollution target; and (3) the optimal trading ratio is not necessarily equal to the delivery coefficient even when it is known with certainty. Intuitively, since the regulator is only required to meet a pollution target on average, she can set the trading ratio and total permit cap such that there will be more pollution when abatement costs are high and less pollution when abatement costs are low. Information on firms’ abatement costs is important in order for the regulator to induce the optimal alignment between pollution level and abatement costs.