# Behavioral Analytics for Myopic Agents

Thursday, October 4, 2018 - 9:00am - 9:45am

Keller 3-180

Philip Kaminsky (University of California, Berkeley)

Many multi-agent systems, including those in supply chain and healthcare settings, have the structure of a single coordinator providing behavioral or financial incentives to a large number of agents. In such settings, two challenges faced by the coordinator are effectively allocating resources from a finite budget, and doing so with a lack of knowledge about the utility functions of the agents. In this work, initially motivated by a healthcare problem, we develop a behavioral analytics approach for solving the coordinator's problem, when the agents make decisions by maximizing utility functions that depend on prior system states, inputs, and other parameters that are initially unknown and subject to partially unknown temporal dynamics. Our behavioral analytics framework involves three step: first, we develop a behavioral model that describes the decision-making process of an agent; second, we use data to estimate behavioral model parameters for each agent and then use these estimates to predict future decisions of each agent; and third, we use the estimated behavioral model parameters to optimize a set of costly incentives to provide to each agent.

We describe a set of tools, models, and approaches that fit into this framework, and that adapt models and incentives as new information is collected by repeating the second and third steps described above. We prove that the incentives computed by this adaptive approach are asymptotically optimal with respect to a given loss function that describes the coordinator's objective. We optimize incentives utilizing a decomposition scheme, where each sub-problem solves the coordinator's problem for a single agent, and the master problem is a pure integer program.

We initially developed and tested this approach in a healthcare setting. We present our results in this setting, and discuss extensions and modifications that would be necessary to use these tools to address problems in a supply chain setting.

Joint work with Anil Aswani, Yonatan Mintz, Mo Zhou, Elena Flowers, Yoshimi Fukuoka, and Ken Goldberg.

We describe a set of tools, models, and approaches that fit into this framework, and that adapt models and incentives as new information is collected by repeating the second and third steps described above. We prove that the incentives computed by this adaptive approach are asymptotically optimal with respect to a given loss function that describes the coordinator's objective. We optimize incentives utilizing a decomposition scheme, where each sub-problem solves the coordinator's problem for a single agent, and the master problem is a pure integer program.

We initially developed and tested this approach in a healthcare setting. We present our results in this setting, and discuss extensions and modifications that would be necessary to use these tools to address problems in a supply chain setting.

Joint work with Anil Aswani, Yonatan Mintz, Mo Zhou, Elena Flowers, Yoshimi Fukuoka, and Ken Goldberg.