About

Babak Heydari is an associate professor of mechanical and industrial engineering at Northeastern University, with affiliations in the Network Science Institute and the School for Public Policy and Urban Affairs. His research focuses on building computational models for group behaviors, such as the formation of social bubbles and changes in risk-mitigating norms and conventions, particularly in the context of epidemics. He is the principal investigator of a $1 million NSF grant awarded for the project “No One Lives in a Bubble: Incorporating Group Dynamics Into Epidemic Models,” which aims to understand how collective behaviors influence the spread of infectious diseases. Heydari's work emphasizes the importance of understanding not just the behavior of the virus, but also how people react to the virus and related policies. He highlights the significance of group-level or collective behavior in epidemic modeling, especially as society moves past initial pandemic shocks. His insights address critical questions about the effectiveness of policies like social distancing and lockdowns, considering how individuals and groups adjust their behaviors in response to the pandemic and policy measures. His research aims to improve epidemic forecasts by integrating human behavior at both individual and group levels, providing more informed guidance for policymakers.

Research topics

• Political Science
• Medicine
• Demographic economics
• Economics
• Virology
• Psychology
• Social psychology
• Public economics
• Economic growth
• Nursing
• Environmental health
• Business

Selected publications

• Core or periphery: Examining where to allocate heterogeneous inventors and the impact on firms' innovation

  Strategic Management Journal · 2025-09-17 · 1 citations

  articleOpen access1st author

  Abstract Research Summary We examine how firms should allocate their inventors to network locations to achieve their best innovation performance. We use an NK model to model the interactions between key factors that influence a firm's innovation: (a) individuals' embeddedness within the firm's network, (b) individuals' heterogeneity across their search distance as well as their adoption (i.e., imitation) propensity, and (c) the complexity of the firm's landscape. We find that in high‐complexity landscapes, firms benefit from allocating low‐adopter agents to the core and high‐adopters to the periphery, thereby promoting more independent search at the core. The opposite is true for low complexity landscapes. We further find that when agent types are unknown, individuals should be allocated based on their adoption propensity versus search distance. Managerial Summary Where should firms locate their heterogeneous inventors within their networks? Simulating a complex innovation landscape, our study focuses on the interaction of three essential factors that may influence firms' innovation: the location of inventors in the core (embedded position) or periphery of a firm's structure, the distance inventors search for solutions, and the adoption propensity with which inventors will imitate their peers' solutions versus being independent and following their own trajectory. Our results show that firms benefit by placing independent searchers who also search far at the core of the firm when facing highly complex problems, and at the periphery otherwise. Though search distance and adoption propensity have substitutive effects on innovation performance, firms should allocate based on inventors' adoption propensity to lower innovation‐performance variance.

  PublisherOA PDFDOI

• Evolution of Innovation in Technology Life cycle: Isolating Innovation Legacy from Quality

  SSRN Electronic Journal · 2025-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Core or Periphery: Examining where to allocate heterogeneous inventors and the impact on firms' innovation

  SSRN Electronic Journal · 2025-01-01 · 1 citations

  preprintOpen access1st authorCorresponding
  PublisherDOI

• Adaptive Network Intervention for Complex Systems: A Hierarchical Graph Reinforcement Learning Approach

  Journal of Computing and Information Science in Engineering · 2025-04-18 · 7 citations

  articleOpen accessSenior author

  Abstract Effective governance and steering of behavior in complex multiagent systems (MAS) are essential for managing system-wide outcomes, particularly in environments where interactions are structured by dynamic networks. In many applications, the goal is to promote pro-social behavior among agents, where network structure plays a pivotal role in shaping these interactions. This article introduces a hierarchical graph reinforcement learning (HGRL) framework that governs such systems through targeted interventions in the network structure. Operating within the constraints of limited managerial authority, the HGRL framework demonstrates superior performance across a range of environmental conditions, outperforming established baseline methods. Our findings highlight the critical influence of agent-to-agent learning (social learning) on system behavior: under low social learning, the HGRL manager preserves cooperation, forming robust core-periphery networks dominated by cooperators. In contrast, high social learning accelerates defection, leading to sparser, chain-like networks. Additionally, the study underscores the importance of the system manager’s authority level in preventing system-wide failures, such as agent rebellion or collapse, positioning HGRL as a powerful tool for dynamic network-based governance.

  PublisherOA PDFDOI

• Generative Network Design for Multi-Agent Coordination: A VAE-Assisted Deep RL Approach

  Journal of Mechanical Design · 2025-05-29 · 3 citations

  articleOpen accessSenior author

  Abstract We introduce a framework that integrates variational autoencoders (VAE) with reinforcement learning (RL) to balance system performance and resource usage in multi-agent systems by dynamically adjusting network structures over time. A key innovation of this method is its capability to handle the vast action space of the network structure. This is achieved by combining Variational Auto-Encoder and Deep Reinforcement Learning to control the latent space encoded from the network structures. The proposed method, evaluated on the modified OpenAI particle environment under various scenarios, not only demonstrates superior performance compared to baselines but also reveals interesting strategies and insights through the learned behaviors.

  PublisherOA PDFDOI

• Network Topology Matters, But Not Always: Mobility Networks in Epidemic Forecasting

  ArXiv.org · 2025-10-22

  preprintOpen accessSenior author

  Short-horizon epidemic forecasts guide near-term staffing, testing, and messaging. Mobility data are now routinely used to improve such forecasts, yet work diverges on whether the volume of mobility or the structure of mobility networks carries the most predictive signal. We study Massachusetts towns (April 2020-April 2021), build a weekly directed mobility network from anonymized smartphone traces, derive dynamic topology measures, and evaluate their out-of-sample value for one-week-ahead COVID-19 forecasts. We compare models that use only macro-level incidence, models that add mobility network features and their interactions with macro incidence, and autoregressive (AR) models that include town-level recent cases. Two results emerge. First, when granular town-level case histories are unavailable, network information (especially interactions between macro incidence and a town's network position) yields large out-of-sample gains (Predict-R2 rising from 0.60 to 0.83-0.89). Second, when town-level case histories are available, AR models capture most short-horizon predictability; adding network features provides only minimal incremental lift (about +0.5 percentage points). Gains from network information are largest during epidemic waves and rising phases, when connectivity and incidence change rapidly. Agent-based simulations reproduce these patterns under controlled dynamics, and a simple analytical decomposition clarifies why network interactions explain a large share of cross-sectional variance when only macro-level counts are available, but much less once recent town-level case histories are included. Together, the results offer a practical decision rule: compute network metrics (and interactions) when local case histories are coarse or delayed; rely primarily on AR baselines when granular cases are timely, using network signals mainly for diagnostic targeting.

  PublisherOA PDFDOI

• Competing for Attention: Explainable Deep Reinforcement Learning for Behavior-Aware Recommendations

  SSRN Electronic Journal · 2025-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Special Issue: Networks and Graphs for Engineering Systems and Design

  Journal of Computing and Information Science in Engineering · 2025-04-16 · 1 citations

  articleOpen access

  In the ever-evolving landscape of engineering, the fusion of network science and graph theories has emerged as a dynamic force, revolutionizing the way we represent, design, model, and optimize complex systems. Networks, defined by nodes and edges, are particularly effective in modeling the interaction and interdependency among individual entities in complex systems. Networks have become the cornerstone for comprehending the intricate relationships underlying a myriad of engineering domains. From transportation networks optimizing urban mobility, power grids ensuring energy efficiency and resilience, and social networks shaping human interactions to biological networks inspiring human-engineered system design, the application of network science and graphs in engineering spans a vast spectrum of disciplines. This special issue is dedicated to promoting the dissemination of knowledge related to complex networks in engineering systems and design and highlighting the latest advances at the intersection of network science, graph theory, and engineering.

  PublisherOA PDFDOI

• The Echo Chambers of Complexity: How Task Complexity Influences Team Groupthink and Individual Exploration

  SSRN Electronic Journal · 2024-01-01 · 2 citations

  preprintOpen accessSenior author
  PublisherDOI

• Platform-Driven Collaboration Patterns: Structural Evolution Over Time and Scale

  IEEE Transactions on Computational Social Systems · 2024-09-24 · 2 citations

  articleSenior author

  Within an increasingly digitalized organizational landscape, this research explores the dynamics of decentralized collaboration, contrasting it with traditional collaboration models. An effective capturing of high-level collaborations (beyond direct messages) is introduced as the network construction methodology including both temporal and content dimensions of user collaborations—an alternating timed interaction (ATI) metric as the first aspect, and a quantitative strategy of thematic similarity as the second aspect. This study validates three hypotheses that collectively underscore the complexities of digital team dynamics within sociotechnical systems. First, it establishes the significant influence of problem context on team structures in work environments. Second, the study reveals specific evolving patterns of team structures on digital platforms concerning team size and problem maturity. Last, it identifies substantial differences in team structure patterns between digital platforms and traditional organizational settings, underscoring the unexplored nature of digital collaboration dynamics. Focusing on Wikipedia's co-creation teams as a representative online platform, this study is instrumental for organizations navigating the digital era by identifying opportunities and challenges for managing information flow. The findings reveal significant collaborative potential and innovation in large online teams: the high speed of knowledge-sharing, numerous subcommunities, and highly decentralized leadership. This study paves the way for platform governors to design strategic interventions, tailored for different problem types, to optimize digital team dynamics and align them to broader organizational goals.

  PublisherDOI

Recent grants

• CAREER: Architecting Products to Balance Innovation and Competition in Business Ecosystems

  NSF · $500k · 2016–2019

• CAREER: Architecting Products to Balance Innovation and Competition in Business Ecosystems

  NSF · $487k · 2018–2022

• EAGER: Hybrid Socio-Technical Teams: A Theoretical Framework for Modeling and Design of Hybrid Networks of Human and Autonomous Agents

  NSF · $230k · 2015–2018

Frequent coauthors

• Ali M. Niknejad

  18 shared

• Mohsen Mosleh

  Massachusetts Institute of Technology

  14 shared

• Ehsan Adabi

  11 shared

• Mounir Bohsali

  11 shared

• Kia Dalili

  Meta (United States)

  10 shared

• Daniel T. O’Brien

  Universidad del Noreste

  10 shared

• Justin De Benedictis-Kessner

  Harvard University

  10 shared

• Alexandra Ciomek

  Boston University

  8 shared

Awards & honors

• National Science Foundation CAREER Award
• Wiley Systems Engineering Journal Best Paper Award
• Best Paper Award, Journal of Systems Engineering (2016)