SSRN Author: Ozan CandoganOzan Candogan SSRN Content
https://privwww.ssrn.com/author=2329383
https://privwww.ssrn.com/rss/en-usThu, 03 Jun 2021 01:22:28 GMTeditor@ssrn.com (Editor)Thu, 03 Jun 2021 01:22:28 GMTwebmaster@ssrn.com (WebMaster)SSRN RSS Generator 1.0REVISION: Near-Optimal Experimental Design for Networks: Independent Block RandomizationMotivated by the prevalence of experimentation in online platforms and social networks, we consider the problem of designing a randomized experiment to assess the effectiveness of a new treatment for a network of users. The outcome of each user depends on the assignment of the treatment to her as well as her neighbors. Given the experiment, the unbiased Horvitz-Thompson estimator is used to estimate the total market effect of the treatment. The decision-maker chooses an optimal joint distribution of randomized assignments of users to treatment and control, in order to minimize the worst-case variance of this estimator. We focus on networks that can be partitioned into communities, where the users in the same community are densely connected, and users from different communities are only loosely connected. In such settings, it is near-optimal to assign all users in the same community to the same variant (treatment or control). The problem of designing the optimal randomized assignments ...
https://privwww.ssrn.com/abstract=3852100
https://privwww.ssrn.com/2030037.htmlWed, 02 Jun 2021 20:17:48 GMTREVISION: Persuasion in Networks: Public Signals and k-CoresWe consider a setting where agents in a social network take binary actions, which exhibit local strategic complementarities. The agents are a priori uninformed about an underlying payoff-relevant state. An information designer wants to maximize the expected number of agents who take action 1, and she can commit to a signaling mechanism which upon the realization of the state sends an informative signal to all the agents. We study the structure and design of the optimal public signaling mechanisms.<br><br>We establish that given a signal realization, the set of agents who take action 1 correspond to a k-core of the underlying network, for some k. Using this we show that the designer’s payoff is an increasing step function of the posterior mean her signals induce. We provide a convex optimization formulation and an algorithm that obtain the optimal information structure whenever the designer’s payoff exhibits this structure. The latter structural property is prevalent, thereby making ...
https://privwww.ssrn.com/abstract=3346144
https://privwww.ssrn.com/2029658.htmlTue, 01 Jun 2021 21:07:16 GMTREVISION: Near-Optimal Experimental Design for Networks: Independent Block RandomizationMotivated by the prevalence of experimentation in online platforms and social networks, we consider the problem of designing a randomized experiment to assess the effectiveness of a new treatment for a network of users. The outcome of each user depends on the assignment of the treatment to her as well as her neighbors. Given the experiment, the unbiased Horvitz-Thompson estimator is used to estimate the total market effect of the treatment. The decision-maker chooses an optimal joint distribution of randomized assignments of users to treatment and control, in order to minimize the worst-case variance of this estimator. We focus on networks that can be partitioned into communities, where the users in the same community are densely connected, and users from different communities are only loosely connected. In such settings, it is near-optimal to assign all users in the same community to the same variant (treatment or control). The problem of designing the optimal randomized assignments ...
https://privwww.ssrn.com/abstract=3852100
https://privwww.ssrn.com/2029350.htmlTue, 01 Jun 2021 15:11:28 GMTNew: Network Inventory Management: Approximate Optimality in Large-Scale SystemsWe consider a discrete time network inventory management problem on a hub-and-spoke network. Each period begins with an initial inventory at each of the nodes in the network, after which the customers (demand) materialize at the nodes. Each customer picks up a unit at the origin node and drops it off at a randomly sampled destination node given by an origin-specific probability distribution. An important task in inventory management of such systems is the periodic physical repositioning of the inventory to avoid stockouts or excess inventory at nodes with unbalanced flows. In our work, we model the above network inventory management problem as an infinite horizon discrete-time discounted Markov Decision Process, and prove the asymptotic optimality of a novel mean-field approximation to the original MDP as the number of spokes becomes large. To compute an approximately optimal policy for the mean-field dynamics, we provide an algorithm whose running time is logarithmic in the desired ...
https://privwww.ssrn.com/abstract=3842817
https://privwww.ssrn.com/2024024.htmlWed, 12 May 2021 11:36:23 GMTREVISION: Optimal Disclosure of Information to a Privately Informed ReceiverWe study information design problems where the designer controls information about a state and the receiver is privately informed about his preferences. The receiver’s action set is general and his preferences depend linearly on the state. We show that to optimally screen the receiver, the designer can use a menu of “laminar partitional” signals. These signals partition the states and send the same non-random message in each partition element. The convex hulls of any two partition elements are such that either one contains the other or they have an empty intersection. Furthermore, each state is either perfectly revealed or lies in an interval in which at most n + 2 different messages are sent, where n is the number of receiver types.<br><br>In the finite action case an optimal menu can be obtained by solving a finite- dimensional convex program. Along the way we shed light on the solutions of optimization problems over distributions subject to a mean-preserving contraction constraint ...
https://privwww.ssrn.com/abstract=3773326
https://privwww.ssrn.com/2011763.htmlMon, 05 Apr 2021 16:43:28 GMTREVISION: Optimal Disclosure of Information to a Privately Informed ReceiverWe study information design problems where the designer controls information about a state and the receiver is privately informed about his preferences. The receiver’s action set is general and his preferences depend linearly on the state. We show that to optimally screen the receiver, the designer can use a menu of “laminar partitional” signals. These signals partition the states and send the same non-random message in each partition element. The convex hulls of any two partition elements are such that either one contains the other or they have an empty intersection. Furthermore, each state is either perfectly revealed or lies in an interval in which at most n + 2 different messages are sent, where n is the number of receiver types.<br><br>In the finite action case an optimal menu can be obtained by solving a finite- dimensional convex program. Along the way we shed light on the solutions of optimization problems over distributions subject to a mean-preserving contraction constraint ...
https://privwww.ssrn.com/abstract=3773326
https://privwww.ssrn.com/2001224.htmlThu, 11 Mar 2021 14:41:35 GMTREVISION: Controlling Epidemic Spread: Reducing Economic Losses with Targeted ClosuresData on population movements can be helpful in designing targeted policy responses to curb epidemic spread. We study a spatial epidemic model, which explicitly accounts for population movements, and propose an optimization framework for obtaining targeted policies that restrict economic activity in different neighborhoods of a city at different levels. We focus on COVID-19 and calibrate our model using the mobile phone data that capture individuals’ movements within New York City (NYC). We show that appropriate targeting achieves a reduction in infections in all neighborhoods while resuming 23.1%–42.4% of the baseline non-teleworkable employment in NYC. By contrast, uniform (city-wide) restriction policies that achieve the same policy goal permit 3.92 to 6.25 times less non-teleworkable employment. Our targeting framework gives policy makers an approach for curbing the spread of epidemics while limiting unemployment.
https://privwww.ssrn.com/abstract=3590621
https://privwww.ssrn.com/1966303.htmlMon, 30 Nov 2020 09:54:26 GMTNew: Information Design in OperationsConsider a set of agents (receivers) whose payoffs depend on an underlying state of the world as well as each other’s actions. Suppose that a designer (sender) commits to a signaling mechanism which reveals payoff-relevant signals to agents when the state is realized. The availability of such signals influences the agents’ actions, and by choosing the signaling mechanism appropriately the designer can induce a desired outcome. Information design studies signaling mechanisms that maximize the pay- off of the designer. In this paper, we first present the classical information design framework and discuss different approaches for characterizing the optimal information structures. We then discuss various applications in the recent operations literature. The applications include signaling (i) content/product quality in networked systems, (ii) product availability in revenue management settings, and (iii) seller quality in two- sided markets. Finally, we present recent work that discusses ...
https://privwww.ssrn.com/abstract=3666252
https://privwww.ssrn.com/1960427.htmlTue, 10 Nov 2020 19:41:00 GMTNew: Social Learning Under Platform Influence: Extreme Consensus and Persistent DisagreementIndividuals increasingly rely on social networking platforms to find information and form opinions. However, there are concerns on whether or how these platforms lead to extremism and polarization, especially since they typically aim to maximize engagement which may not align with other social objectives. In this work, we introduce an opinion dynamics model where agents are connected in a social network, and repeatedly update their opinions based on the content shown to them by the platform's personalized recommendation and their neighbors' opinions. We prove that agents always converge to some limiting opinion, which can be categorized into two groups: extreme consensus where all agents agree on an extreme opinion, and persistent disagreement where agents disagree. Extreme consensus is more likely when the platform's influence is weak and connections between agents with differing opinions are strong. The platform increases the extremism of opinions when its influence is either weak ...
https://privwww.ssrn.com/abstract=3675712
https://privwww.ssrn.com/1947431.htmlFri, 02 Oct 2020 14:47:21 GMTREVISION: On Information Design with SpilloversAn information designer has access to a set of experiments and decides which of these to assign to each of the agents in a directed network. The network encodes informational spillovers: an agent has access to the experiments assigned to her, as well as to those assigned to any other agent who has a directed path to her. We establish that the designer's problem in any network can be reduced to an equivalent problem in a directed acyclic network. We show that when in the latter network each agent follows at most one other agent (i.e., each node has in-degree at most one), the optimal information structure can be obtained in a tractable way. The problem becomes intractable if some agents follow multiple other agents. Thus, qualitatively, following multiple information sources is what makes information design problems intractable in the presence of spillovers. We also study a voting game with binary actions in the presence of spillovers. We show that when the followers are more ...
https://privwww.ssrn.com/abstract=3537289
https://privwww.ssrn.com/1919391.htmlFri, 10 Jul 2020 08:27:31 GMTREVISION: Persuasion in Networks: Public Signals and k-CoresWe consider a setting where agents in a social network take binary actions, which exhibit local strategic complementarities. The agents are a priori uninformed about an underlying payoff-relevant state. An information designer wants to maximize the expected number of agents who take action 1, and she can commit to a signaling mechanism which upon the realization of the state sends an informative signal to all the agents. We study the structure and design of the optimal public signaling mechanisms.<br><br>We establish that to find an optimal mechanism it suffices to restrict attention to mechanisms where (i) the possible signal realizations correspond to the k-cores of the underlying network, (ii) once the signal realization is k, the agents in the k-core take action 1. Using this observation, we obtain a convex optimization relaxation of the problem of the information designer, and establish that using an optimal solution of this problem together with an algorithm we provide, the ...
https://privwww.ssrn.com/abstract=3346144
https://privwww.ssrn.com/1908258.htmlFri, 12 Jun 2020 08:27:48 GMT