Coordination in complex systems: increasing efficiency in disaster mitigation and response

62 Int. J. Emergency Management, Vol. 2, Nos. 1–2, 2004

Coordination in complex systems: increasing efficiency in disaster mitigation and response

Louise K. Comfort*, Mark Dunn, David Johnson, Robert Skertich and Adam Zagorecki Interactive, Intelligent, Spatial Information System (IISIS) Project University Centre for Social and Urban Research, Room 409 University of Pittsburgh, Pittsburgh, PA 15260, USA E-mail: [email protected] E-mail: [email protected] E-mail: [email protected] E-mail: [email protected] E-mail: [email protected] *Corresponding author

Abstract: Coordination in multi-organisational settings is extraordinarily difficult to achieve. This article examines the problem of inter-organisational coordination in the context of public administration theory and practice. The authors present the concept of complex adaptive systems as a theoretical framework that explains the dynamic processes involved in achieving coordinated action among multiple organisations to manage complex technical operations in environments vulnerable to risk. They argue that coordination may be achieved more easily with the appropriate design of a socio-technical system, that is, a system that supports the exchange of critical information among technical and organisational entities to improve performance in both. The goal is to design a decision support system that uses information technology to enhance the capacity of multiple organisations to adapt their actions reciprocally to changing conditions of risk, enabling the set of organisations to manage risk more effectively and efficiently for the community as a whole. The authors present the design and initial findings from a trial demonstration to implement a prototype interactive, intelligent, spatial information system in the Pittsburgh Metropolitan Region.

Keywords: coordination; complex adaptive systems; self-organisation; socio-technical systems.

Reference to this paper should be made as follows: Comfort, L.K., Dunn, M., Johnson, D., Skertich, R. and Zagorecki, A. (2004) ‘Coordination in complex systems: increasing efficiency in disaster mitigation and response’, Int. J. Emergency Management, Vol. 2, Nos. 1–2, pp.62–80.

Biographical notes: Louise K. Comfort is a Professor of Public and International Affairs at the University of Pittsburgh. She teaches in the field of information policy, organisational theory and policy design, and is the Principal Investigator, Interactive, Intelligent, Spatial Information System (IISIS) Project, 1994-present: http://www.iisis.pitt.edu. She has served as Principal Investigator/Project Coordinator/Co-Principal Investigator/Team Member on 43 funded research projects. She has conducted field research on information processes in disaster operations following 14 earthquakes in ten countries. She holds degrees in political science from Macalester College (B.A.); University of California, Berkeley (M.A.), and Yale University (Ph.D.). Her publications include research on complex systems.

Copyright © 2004 Inderscience Enterprises Ltd.

http://www.iisis.pitt.edu/

Coordination in complex systems 63

Mark Dunn is an experienced Database Administrator and Developer for the IISIS Project. He received his Master of Policy and Management (MPM ‘97) degree from the H. John Heinz III School of Public Policy and Management, Carnegie Mellon University, and his Bachelor’s Degree (BA ‘94) in Public Administration from the University of Pittsburgh. Mark joined the IISIS Team in 1996 and has worked for several private sector firms since 1998. His technical skills include expertise in UNIX, Linux, and Windows operating systems, Shell scripting (korn, C, bash), PL/SQL, HTML languages, and extensive experience with Oracle software.

David Johnson is a doctoral candidate in the Graduate School of Public and International Affairs (GSPIA), University of Pittsburgh. He has 26 years of experience as a Public Safety Practitioner and is currently employed as a Planner in the Allegheny County Department of Emergency Services. He is a Counter Terrorism Instructor for the Pennsylvania Emergency Management Agency and a Deputy Commander for the Disaster Medical Assistance Team Pennsylvania-1. He has a BS in Biopsychology from Juniata College, and an MPA degree from GSPIA. He has served a member of the IISIS research team for five years.

Robert Skertich’s lifelong dedication to public service includes 25 years in public safety management within the academic, government and private industries. A past educator, he has earned numerous academic awards and honors for his work. Mr. Skertich is currently the Chief Operations Officer for the American Red Cross, Southwestern Pennsylvania Chapter. In 2002, he earned his Masters of Public Policy Management from the University of Pittsburgh, Graduate School of Public and International Affairs, where he is completing his doctoral studies in public administration and policy. Rob has served as a graduate student researcher on the IISIS project from 2001 to the present.

Adam Zagorecki is a doctoral student at the School of Information Sciences, University of Pittsburgh, Pittsburgh, PA 15260. He is a member of Decision Systems Laboratory and takes an active role in Interactive, Intelligent, spatial Information System (IISIS) Project. His research interests include modeling uncertainty, graphical models, machine learning and agent-based simulation with particular accent on real-life applications. He holds a M.S. degree in Computer Science from Bialystok University of Technology, Bialystok, Poland.

1 Introduction: coordination as an elusive goal

Administrators and managers in public, private and nonprofit organisations repeatedly call for improved coordination to maximise effort and minimise costs in achieving shared goals for public safety, economic development, healthcare or other vital community services. With solemn commitment and brave rhetoric, many of these organisations seek ways of sharing resources, information, and tasks in order to achieve a common approach to meeting public needs. Yet, as the situation becomes more complex, and the interactions among the organisations become more interdependent, these bold efforts to increase coordination to achieve a common vision for the community frequently fail. Why is coordination so admired in theory, but so difficult to achieve in practice?

64 L.K. Comfort, M. Dunn, D. Johnson, R. Skertich and A. Zagorecki

The barriers to coordination may lie more in the structure of organisations seeking a common approach to action than in any misconstruction of the goal itself. Coordination requires an intensive exchange of information among all participating organisations in order to confirm the specific details of time, costs, and actions necessary to achieve a shared goal. Yet, as the information requirements for inter-organisational coordination increase, the cognitive capacity of human decision makers to process the expanding amounts of information decreases (Miller, 1967; Simon, 1981). Without timely, accurate information to allow the participating organisations to adapt their respective actions to accommodate changing conditions and shifting priorities, efforts to achieve coordinated performance inevitably fail (LaPorte, 1975).

The challenge for administrators and managers who seek coordination in inter-organisational performance lies in recognising that it is fundamentally a voluntary activity sustained by a clearly articulated goal, a shared knowledge base, and a set of systematic information search, exchange and feedback processes. These information processes, linked to a shared knowledge base, allow the participants to learn as they engage in interrelated actions and to adjust their actions accordingly, based upon the most current assessment of their operating environment. The information processes drive the dynamics of organisational adaptation and self organising behaviour, as participating organisations reciprocally adjust their actions to changing conditions in their efforts to achieve a shared goal. Self-organising behaviour among a set of organisations leads to the mutual adaptation and reciprocity that represents coordination in practice.

Creating an environment that supports self-organisation, however, is not easy. It requires scalable interactions among multiple organisations operating at different levels of authority, responsibility and specificity of tasks. Such complex interactions are beyond the capacity of individual memory banks and problem solving ability. They require the technical assistance of an information infrastructure that is designed to build support for organisational policy makers operating in dynamic environments.

Such an information infrastructure would link organisations, individual policy makers, groups of clientele, and computers into a distinct socio-technical system that uses the flexibility of current information technology to support adaptive behaviour by individuals and organisations (Coakes, Willis and Clarke, 2002). With increasing complexity in the organisational, technical, and physical environments that trigger emergencies, higher social, economic, and political costs of failure in such events, and greater access to robust hardware and software for practicing managers, the need to develop a more effective approach for risk reduction and response is imperative. The challenge is to design a socio-technical system (STS) that will capture accurately the rapidly changing dynamics of an emergency environment and support human managers in adapting their performance to meet emerging needs.

A significant body of research exists on STS, but much of this research has addressed problems faced by business organisations, operating as single entities and focused on internal management issues (Coovert and Thompson, 2001). Developing a working STS in the public arena, specifically for emergency management, is a much more complex undertaking. Yet, the same general principles apply, as STSs in both public and private arenas seek to enhance the performance of human actors by the skillful design of technical support for knowledge management. Successfully implemented, a STS optimises the performance of both technical and organisational systems, as informed human managers observe anomalies in the operation of technical systems, and initiate


interventions that maintain continuity in operations, despite disruptions or failure in local or specific areas.

An effective STS in emergency management requires the identification of linkages between the technical systems that are central to daily operations for a city or region, such as communications, electrical power, transportation, water, gas, and sewage distribution systems, and the public, private, and nonprofit organisations that manage them. The knowledge requirements also include different levels of legal responsibility for action, different levels of resources available for action, and different levels of training and experience among the personnel who need to coordinate actions to minimise risk in a shared, public arena. Further, a STS for emergency management needs to capture the changing exposure to risk for a given region, and calibrate that level of risk against the resources and time available to manage it. The goal of a STS in emergency management is to produce a system that is capable of adapting to changes in the risk environment and modifying its behaviour and allocation of attention and resources accordingly (Coakes, Willis and Clarke, 2002; Comfort, 1999)

In practice, a STS represents an interacting set of individuals, organisations, and technical entities that are capable of adjusting their behaviour reciprocally to one another and to their operating environment in order to achieve a shared goal of improved performance (Comfort, 1994; Comfort, 1999). The technical components extend the knowledge base, memory and reasoning capacity of the individuals and organisations that participate in the system. The individuals and organisations, in turn, monitor the performance of the technical components to ensure that they are functioning as expected and provide effective decision support for organisational and inter-organisational problem solving processes under diverse conditions.

The crux of the long-standing problem of achieving coordination in inter-organisational performance (Caiden and Wildavsky, 1974,pp.277–279; Wilson, 1989,pp.268–274; Chambers, 1974,pp.25–26) is creating an information infrastructure that is sufficiently flexible to manage the dynamic exchange of information among the participating entities in an inter-organisational system, but sufficiently ordered to ensure that the relevant information gets to the responsible parties in valid format and in time to support effective action.

In this analysis, we undertake four tasks:

1 We examine briefly the concept of complex adaptive systems as a theoretical framework for analysing the dynamics involved in inter-organisational coordination.

2 We present the research design and methods used in an investigation of uses of information technology to improve coordination among a set of practicing public agencies.

3 We present initial findings from a trial demonstration project to design and implement a socio-technical system for disaster mitigation and management in the Pittsburgh Metropolitan Region.

4 We draw initial conclusions from these findings regarding the requirements for, and chief obstacles to, achieving coordination among public agencies responsible for disaster reduction and management in their communities.


2 Theoretical background

The problem of coordination in organisational performance has been perceived as a conflict between order and flexibility. Classical theorists in public administration have addressed problems of attaining order, stability and accountability in single organisations (Taylor, 1967; Kaufman, 1960; Fesler and Kettl, 1991). Others have acknowledged the problems of public organisations operating in dynamic environments (Waldo, 1971; Rosenau, 1997; Ostrom, 1998), and have noted the importance of flexibility, but also the high cost of attaining it (LaPorte and Consolini, 1991; Chisholm, 1989; Roberts, 1993; Rochlin, 1993). The difficulty in each of these approaches has been to balance order with flexibility, cost with time, and efficiency with effectiveness in the practicing world of public managers.

The concept of complex adaptive systems, coupled with an appropriate design for an information infrastructure to support its application to practicing organisations, offers a theoretical framework that links order to flexibility on a continuum in organisational performance (Kauffman, 1993; Gell-Mann, 1994; Holland, 1995). This framework represents a productive approach for understanding the dynamic processes involved in the performance of public organisations in rapidly evolving environments (Axelrod, 1984; Cohen, 1986; Comfort, 1999; Axelrod and Cohen, 1999). It acknowledges self-organising behaviour based upon incoming information (Kauffman, 1993) as the driving force underlying the dynamics of complex adaptive systems. In this concept, the agents1 involved are able to adapt their actions to changes in their environment based upon observation or incoming information (Holland, 1975; 1995). For the process of self-organisation to function across a set of individuals or organisations, there needs to be a means of transmitting and processing the information so that the individuals (or agents) may use it as a basis of decision regarding their actions. The interacting components represent a complex system, with recurring patterns of information search, exchange, and adaptive behaviour.

The dynamics of complex adaptive systems are ever-changing. Stuart Kaufmann, a research biologist, states that all systems operate on a continuum ranging from chaos to order, and systems at either end of the continuum tend to move toward the centre. That is, in a chaotic situation, a system will move toward order. In an orderly situation, a system will move toward chaos. In the centre of the continuum between chaos and order lies a narrow region that Kauffman, 1993, 174,pp.208–227) identifies as the ‘edge of chaos,’ where there is sufficient structure to hold and exchange information, and sufficient flexibility to adapt to changing conditions. It is in this narrow region, the ‘edge of chaos,’ that organisations and systems are able to make the most creative responses to changing conditions in their operating environments.

The requirements for action in complex social systems depend on the extent and effectiveness of information and communication processes operating within the inter-organisational system, and its ensuing capacity to foster collective learning and adaptation. Without the capacity for continuing assessments of their current operating conditions and environment, organisations are likely to miss critical points at which changes in these conditions or in the performance of other organisations necessitate adjustments in their own performance. When the intelligence-gathering function falters, the capacity of individual units in the system to ‘make sense’ of their operating environment fails (Weick, 1993). This capacity for sense-making is essential for action. Without it, individuals and organisations frequently ‘lock out’ information that is


inconsistent with their prior assessment (Cohen and Levinthal, 1990,p.136), and are unable to integrate relevant new information into their previous knowledge base. Lacking relevant information to form a system-wide perspective, individual units make separate decisions that, while appropriate for the individual unit, may counter or conflict with the system-wide goal and prove adverse to other units within the system. Under these conditions, relationships among the units within the system become brittle, and tend to shatter under pressure of urgent, time-sensitive events (Comfort, 1999).

Advances in information technology offer a means of addressing problems of communication and coordination that is central to improving performance within and among multiple organisations operating in complex environments. Coordination requires information scaled to different levels of responsibility and action, which must be updated and re-validated in a rapidly changing environment. Integrating information technology directly into the administrative systems for mitigating and managing risk in dynamic environments creates a distinct, socio-technical system that uses the technical functions of information technology to facilitate the information functions required of a multi-organisational system. Such systems can function only with well-trained personnel who have both the technical skills for designing and implementing a distributed knowledge base, and the organisational skills for analysing the degrees of risk to the system from varying environmental conditions and mobilising inter-organisational resources in coherent strategies to reduce that risk.

Developing a systematic approach to studying the dynamics of socio-technical systems and their contribution to innovative organisational performance requires identifying a field environment in which socio-technical systems can be observed and, with care, simulated for study. Effective mobilisation of response to extreme events on a large scale offers a challenging environment in which to study the dynamics of multi-organisational coordination. Rapid response to extreme events is one of the least understood problems in administrative practice. It requires the rapid search, exchange, and absorption of valid information regarding sudden, damaging events transmitted through a network of organisations that crosses disciplinary, jurisdictional and organisational boundaries. It requires pre-event planning among organisations to identify what information will be required and how this information may be accessed. It entails the rapid comprehension of danger that, under ordinary circumstances, is unimaginable, and requires the capacity to use that powerful insight to anticipate the spread of risk through an interdependent community and to devise actions that will interrupt or limit the risk. It means discovering the logic that will govern uncertainty in technical and organisational performance, caused by the event (Comfort, 1989). This is an inference process that functions more through recognition of signals and symbols (Feldman and March, 1981), than on rule-based reasoning (Hayes-Roth, Waterman and Lenat, 1983).

While the collapse of organisational capacity to act under extreme conditions has been vividly documented in actual cases (Weick, 1993; Carley and Harrald, 1997; Comfort, 1999), the opposite phenomenon, the design and development of communities capable of innovative and responsible performance under threat of extreme danger has not been studied systematically. There has been no systematic effort to model the effects of the rapid spread of information regarding risk on the performance of communities under threat, or to estimate the economic costs and social benefits of making the investment in information technology and organisational training that would be necessary to achieve ‘reliable performance’ in extreme events. In this article, we summarise the


research design and implementation process for a trial demonstration of a prototype interactive, intelligent, spatial information system (IISIS) to facilitate coordination among local governments for reduction of, and response to, hazardous materials incidents in the Pittsburgh Metropolitan Region. This approach acknowledges the organisational and policy processes that contribute to change, learning, and innovation in dynamic environments (Peitgen, Saupe and Jurgens, 1992; Argyris, 1993; Comfort, 1994), but considers these processes in a specific context, that of system-wide response to a massive event.

3 Research design

The trial demonstration was designed to address three interrelated problems that characterise decision making in complex environments. These problems are exacerbated in disaster environments, when time and resources are scarce and human lives and properties are at risk:

1 Recurring failure in coordination among multiple organisations participating in environments of organised complexity is rooted in the limited capacity of humans to search, process, understand and act on information needed for collective problem solving (Miller, 1969; Simon, 1981; Simon, 2000).

2 This condition creates the need to develop increased capacity for self-organising performance among organisations operating on a system-wide scale to counter failed coordination (Comfort, 1994).

3 Self-organising action may lead to different strategies for solving the same problem by different organisations, resulting in inefficient use of scarce resources and time for the community under threat.

Evaluation of innovative problem solving strategies devised by organisations under different conditions of threat, urgency, and access to information becomes essential to integrate effective strategies into a coherent program of coordinated action for the community.

Information search, processing, and exchange are critical factors that affect decision making in all three problems. Lack of information leads to failure in coordination, while its timely communication contributes to informed adaptation among organisations under threat. Coordination depends upon the rapid exchange and comprehension of information among human managers in multiple organisations and jurisdictions operating under dynamic conditions. Evaluating different strategies for enhancing information search, processing and exchange among organisations creates the possibility of increasing coordination through feedback processes and informed choice among multiple managers operating in different locations at risk. Increased coordination in disaster response operations leads to decreased losses for the community.

The design, delivery, and measurement of effective information processes on a community-wide scale are not trivial. The difficulty increases under the urgency and complexity imposed by extreme contexts. While recent research on computer-aided decision making has focused on increasing the capacity of individuals or small groups to make more informed decisions (Hoschka, 1996), there has been no previous exploration of the extent to which computer-aided decision support can increase the capacity for


coordinated action among sets of organisations working at different levels of responsibility in multiple jurisdictions to achieve a common goal.

Research questions investigating the performance of socio-technical systems in extreme events focus on the interaction not only between different organisational units involved in the operation of the system, but also on the interactions between the organisational units and the technical components of the system. Consequently, an appropriate research design must examine the extent to which the technical functions of information technology are fitted to the organisational requirements for decision-making and action.

Different approaches have been used to study the problem of coordination in multi-organisational systems. Early approaches relied primarily on redundancy (Landau, 1969; Landau, 1991) or organisational innovations for sharing information and reducing barriers to communication among participating units; (Chisholm, 1989; Feldman, 1989). The use of redundancy, standard practice in emergency response organisations, has proven costly in an era of ‘cutback government,’ and organisations that had adopted this means are now searching for improved strategies. Innovations in organisational structure and information processes have had the welcome effect of legitimising improved personal communications among members of large organisations, but have proven regrettably unreliable in practice.

A second approach involves the simulation of organisational processes using a system of networked computers (Carley, 1998; Carley and Lee, 1998). This approach shows significant promise, but its validity rests upon obtaining sufficient data from the organisations under study to model the organisational processes reliably. A third approach is to conduct a trial demonstration of a prototype decision support system (DSS) with actual organisations that share a common goal and have interrelated responsibilities for attaining that goal. We adopted this approach at the University of Pittsburgh to conduct a trial demonstration of a prototype interactive, intelligent spatial information system (IISIS) in the Pittsburgh Metropolitan Region in 2001. This approach is not simple, as it requires gaining the cooperation of practicing agencies to participate in a trial demonstration of the prototype DSS, but it has the distinct advantage of providing a realistic environment in which to test assumptions about the capacity of practicing organisations to learn new skills and to engage in system-wide efforts to improve performance.

4 The IISIS prototype: a trial demonstration in the Pittsburgh metropolitan region

With advice and counsel from practicing emergency managers, an interdisciplinary research team at the University of Pittsburgh undertook a trial demonstration of a prototype interactive, intelligent, spatial information system (IISIS) involving six entities exposed to similar types of risk in the Pittsburgh Metropolitan Region. The IISIS is a decision support system that adapts advanced information technologies to support increased coordination among multiple organisations at different jurisdictional levels engaged in risk reduction and response operations. The problem of coordination is especially critical for emergency managers operating at the community level where resources and training are often limited.


The prototype IISIS links three types of information technology to create an event-specific knowledge base that can provide timely, valid information to practicing managers as conditions change and demands for coordinated action increase in a complex, dynamic event. The three technologies include:

1 Interactive communication via both internet and secure intranet networks.

2 GIS and remote sensing imagery to provide graphic representation of changes in the area.

3 Intelligent reasoning by the computer to provide estimates of known losses or probabilities of likely consequences that could result from the event, e.g. fires following accidents, failures in transportation networks, hazardous materials releases, or public health needs.

The trial demonstration was designed to assess the capacity of the IISIS prototype to assist practicing managers in coordinating their performance at the community level.

The trial demonstration included six entities operating at three jurisdictional levels of local government in the Pittsburgh Metropolitan Region. The jurisdictional levels were:

1 University level, represented by the University of Pittsburgh that functions as a legal jurisdiction with a daytime population of approximately 32,000.

2 Municipal level, represented by three municipalities of varying sizes:

• City of Pittsburgh, population, 340,000

• Municipality of Penn Hills, population, 46,000

• Wilkins Township, population, 8,000

3 County level, represented by Allegheny County, population, 1.26 million and the American Red Cross, Southwestern Pennsylvania Chapter in its services to Allegheny County.

Representatives of emergency response organisations in each entity agreed to participate as partners in this project, and to review the functionality of the prototype IISIS in terms of the needs of practicing emergency managers.

Within each jurisdiction, public, private and nonprofit organisations need to coordinate their actions for timely, effective mitigation and response to disaster. This set of actions must also be coordinated simultaneously between the jurisdictions to create a dynamic emergency response system for a threatening event. Failure in any one organisation triggers failure in other processes and systems, whereas effective coordination across organisations enables the response system to recover from a specific breakdown more effectively and to reduce the probability of failure spreading throughout the community. The IISIS prototype is designed to provide timely, accurate decision support for practicing emergency managers, thereby increasing their capacity for coordination in emergency response.


5 Goals of the IISIS project

The IISIS prototype has three fundamental goals:

1 It seeks to reduce risk for organisations operating at the community level. It is an innovative, socio-technical system designed to enhance the capacity for risk assessment and disaster management of local organisations. That is, it uses several types of information technology to support a flexible organisational design that provides decision support to managers who are responsible for emergency services in public, private and nonprofit organisations at the community level. These managers include:

• Responders such as police, fire and emergency medical services for public organizations.

• Emergency coordinators for hospitals, schools, libraries, museums and nonprofit organisations that serve dependent populations.

• As well as emergency coordinators for private organisations such as banks, corporations, and small businesses.

2 The IISIS prototype seeks to improve performance for practicing managers. It includes a component of training and education for managers in public, nonprofit and private organisations engaged in disaster mitigation and response.

3 Once established and operational in a local community, an IISIS would serve as a resource centre and demonstration model for other organisations and communities seeking to improve their capacity for disaster management.

6 System design

The basic design of the IISIS prototype integrates three types of information technology to facilitate decision-making in complex, dynamic environments. In recent years, the interdisciplinary research team at the University of Pittsburgh has developed a prototype software system to support the complex decision making processes in emergency management. Essential elements of the system include the ability to support:

• real-time communication across departments, agencies and jurisdictions

• real-time access to distributed databases that can be used to define the nature, scope and potential impact of emergency situations

• real-time access to GIS systems to map critical events and responses

• rapid assessment of probable risk to different groups and different sections of the community under threat, using computer calculations that compare incoming information about the threat to stored information about the population, infrastructure, and vulnerability of the community

• calculation of time, cost, and consequences of alternative strategies for action, based upon priorities and risks specified by emergency managers.


For example, an IISIS in operation at the University of Pittsburgh would provide simultaneous transmission of incident status reports regarding a threatening event to all managers of University departments who are designated as first responders on campus. In turn, these managers would access relevant databases at the University that would enable them to assess the likely impact of the threat upon University personnel and campus facilities. They would also be able to access maps of the campus, with detailed floor plans of campus buildings, and match the physical location of the threat to information stored in departmental databases about equipment, lifelines, facilities and personnel who may be at risk. Using both previously defined rules for emergency response and models that calculate the probability of risk to personnel and property under changing conditions, the IISIS would enable practicing managers to make timely, informed decisions regarding alternative strategies of action to bring the threat under control and reduce the risk of further danger or destruction.

The actual knowledge base to support the decision process in emergency management needs to be developed for the operating context of the participating organisations, their conditions, and constraints. In this process, IISIS staff members use the basic system design for the prototype, but fit the particular characteristics of the specific jurisdiction to that design. This process is facilitated by recognition of the need for national standards in the development of databases for emergency management. The IISIS prototype, for example, uses the basic categories of the Incident Management System, first developed in California as the Incident Command System and now widely used by emergency response agencies across the nation, including the Federal Emergency Management Agency (FEMA) in the Department of Homeland Security. It also follows the standards of the National Spatial Data Infrastructure (NSDI) in formatting spatial data.

The concept underlying IISIS is that of a self-organising, learning system in which individuals, computers, and organisations facilitate the exchange and analysis of timely, accurate information to improve the capacity for human decision makers to make more informed decisions regarding shared problems. The function of IISIS is to simplify and coordinate the inherently complex process of emergency management, not to create a process that is itself complex. This design can be developed at city, county, state, national and international levels, as sets of organisations confront policy problems that cross jurisdictional boundaries.

In the demonstration project year, the research team focused on two important aspects of the IISIS prototype:

1 Extended the prototype from the University of Pittsburgh to emergency response departments in the City of Pittsburgh, Municipality of Penn Hills, Township of Wilkins, Allegheny County, and the Southwestern Pennsylvania Chapter of the US Red Cross. This trial demonstration adapted the IISIS prototype to support interactions among university, municipality and county agencies in a larger, more complex and diverse system of emergency preparedness, mitigation and response.

2 The research team worked intensively on the development of rule-based and probabilistic models of intelligent reasoning for the computer. The rule-based models build on the official rules and standard procedures that have been accepted by emergency managers and their organisations.

Practicing managers may use these models to check incoming information against standard procedures to confirm that their actions are consistent with existing procedures


or to show where they are not. Probabilistic models use the computer to calculate the likely impact of a given event upon the population or built infrastructure of a community, under a range of different conditions. Both types of models become increasingly important to timely, effective decision support as a community environment increases in size, complexity and exposure to threat.

7 Evaluation of expected outcomes

Outcomes expected from the proposed project included a demonstrated increase in efficiency in first response to, as well as effectiveness in the mitigation of, disaster at the community level. The performance of this trial demonstration of the IISIS prototype was tested in a simulated operations exercise on December 4, 2001. This exercise involved problems that required coordination of actions and shared knowledge within and among three levels of local government in the Pittsburgh Metropolitan Region:

1 university

2 city

3 county.

The IISIS prototype differs from other software programs in emergency management in that it combines current theory in organisational design, feedback and learning with experienced judgment from practicing managers who have worked in dynamic disaster environments. Consequently, the functions of the software directly address the stated needs of emergency managers. For example, when asked what were their most critical needs for decision support, practicing managers identified four needs that recur in every disaster event.2

1 They wished to have a means of assessing the vulnerability of their communities to specific types of hazards. This function is served by the initial design of the knowledge base for that community, using classification methods that are accepted practice in emergency management.

2 Practicing managers wished to assess the likely damage of an incident quickly and to monitor the actual costs on a daily basis. These functions are performed by the IISIS components on Damage Assessment and Finance.

3 They wished to access the emergency plans and resources available at upper and lower levels of jurisdictional assistance as the response system evolved for a given disaster. These functions are performed by the IISIS components on Planning and Logistics.

4 They wished to calculate the likely consequences of a particular threat for a given segment of the population, a given section of the affected geographic region, or a distinct set of lifeline services within the threatened region. This function is being addressed through the development of the intelligent reasoning component with its calculation of probabilities.

A major objective of the trial demonstration project was to test the reliability of these and other functions of the IISIS prototype in a simulated setting with practicing managers.


Their feedback was considered invaluable to making adjustments in the performance of the IISIS prototype prior to disseminating findings to a wider audience of public, private and nonprofit managers interested in risk reduction.

The intelligent reasoning components of the IISIS prototype present new features in decision support for emergency managers that are not available in existing software programs. For example, the ASK model uses Bayesian statistics to calculate probabilities that would likely be assigned to specific response actions by experienced emergency managers in risk environments under a shifting mix of conditions. Further, the development of the shared knowledge base to support the functioning IISIS is a central component of the system. The consulting and training services essential to assist a local community, with different types of organisations that have different needs for information, will be included as part of the IISIS product. An important aspect of maintaining the currency and effectiveness of the IISIS will be a continuing research relationship with the University to update the IISIS functions on the basis of new situations and conditions reported from the field. For more information and updates on the progress of this research, please check the project’s web page, http://www.iisis.pitt.edu/

8 Trial demonstration project in the Pittsburgh metropolitan region

The simulated operations exercise held on December 4, 2001 served as a test of performance for the IISIS prototype. In addition, findings from a set of expert interviews with responsible emergency managers in each of the six participating entities3 identified five critical issues that affect performance in inter-organisational emergency response operations. These are issues that, with careful design, may be addressed more effectively through the use of a socio-technical DSS, such as the IISIS prototype. The issues are:

• Intermittency in the assessment of risk both within and among the six entities.

• Lack of timely information regarding the likely impact of an impending threat upon critical infrastructure in the community, and the capacity to anticipate the likely consequences for different populations in the community.

• Loss of information when an emergency escalates from one jurisdictional level to the next.

• Lack of real-time monitoring of threatening events, and the capacity to transmit this information between field operations units and Emergency Operations Centres.

• Lack of a systematic means of monitoring multiple sites simultaneously in order to provide practicing managers with a comprehensive profile of the evolving disaster in a region-wide event.

Each of these issues bears substantive discussion regarding its impact upon the capacity of a community to assess and manage its own risk. Each also indicates a gap in the flow of information through the designated organisational response system designed to mitigate and respond to threats to the safety and stability of the community. The set of issues identified by the practicing managers indicates the critical importance of timely, accurate information distributed reliably throughout the inter-organisational response system to enable the multiple actors/agents to adapt their performance accordingly as


conditions change. Failure on any one issue would generate different consequences for the performance of the system, but failure on the full set of issues occurring simultaneously could lead to a cumulative breakdown of the entire system, shifting it toward the chaotic end of the continuum in emergency response. For example, the inability of emergency response organisations to mobilise effectively in the first hours after the severe earthquake on January 17, 1995 in the Hanshin-Awajii Metropolitan Region of Japan provides a vivid illustration of the cumulative effects of inadequate information processes upon coordination in complex systems (Comfort, 1999,Chapter 8).

The set of issues identified by the emergency managers indicates different points of failure within a complex emergency response system from a regional perspective. The first issue, intermittency, points to the importance as well as difficulty of maintaining a common focus and timing among the set of emergency managers in mobilising coordinated response. If the different entities or jurisdictions are operating at different levels of interest, attention, training and assessment of vulnerability to hazards, their capacity for coordinated action is seriously impaired. Worse, there appears to be a downward spiral in the effort to maintain a common focus. If one organisation initiates a strong effort to analyse hazards in the community, but this effort is not matched by other responsible organisations, the first organisation is unable to achieve its goal due to its interdependence with other organisations in the region. Dismayed, it may drop its efforts, just when another organisation may begin. If there is not a regular, systematic process to update information, training, equipment and skills by all participating system units within a common time frame, the effect of ‘start and stop’ efforts to improve capacity for emergency response reduces the motivation for all units and has a strongly negative effect upon the system.4

This issue may be effectively addressed by the interactive communication function of the IISIS prototype. Modeled after the Incident Management System in common use by emergency response agencies, the communication function of the IISIS prototype enables simultaneous transmission of incoming information to each of the relevant players in the community disaster response system.5 With its capacity for regular updates from the field and system-wide transmission of incoming information, the prototype IISIS would support the maintenance of a regular schedule for emergency preparedness and mitigation activities within a jurisdiction or metropolitan region.

The second issue, lack of timely information about risk, compounds the negative effects of the first. It addresses the need not only for systematic development of a valid knowledge base, but also the capacity to identify, analyse and estimate the likelihood of risk to the community on a continuing basis. The distributed knowledge base of the IISIS prototype and its modeling capacity for certain types of risk represent a means of estimating the probabilities of risk for different constituencies within a jurisdiction and different districts of the community.

Third, the inevitable loss of information when an evolving emergency response system moves from one level of jurisdiction to the next can be minimised significantly by the Operations and Incident Log functions of the IISIS prototype. The Operations function tracks the course of all actions initiated, continued, interrupted or delayed, and completed by site and by response system, so that emergency managers can follow the sequence of events for the whole system, or, specifically, the information on each site. The Incident Log function keeps a chronological record of all actions taken during disaster operations for any given event. It forms a disaster-specific knowledge base that


allows emergency managers to review their actions taken and consider alternative methods of response. It is especially relevant for documentation of events after the disaster, and in training personnel for response to future disasters.

Fourth, the gap in information between field units and command centres is a continuing problem in emergency operations. By facilitating not only the exchange of information between field units and Emergency Operations Centres, but also access to relevant types of knowledge from a distributed knowledge base, the IISIS prototype will enable more efficient operations in both time and cost.

Finally, the most difficult problem for emergency managers is to maintain a comprehensive profile of the entire range of disaster operations, while simultaneously being able to get detailed information on specific sites or incidents. This function requires that they operate on at least at two levels of abstraction simultaneously, or be able to move quickly from a general profile of the disaster to a specific incident at a given location and back again to the general profile. Shifting from one level of aggregation of information to the next, while maintaining the appropriate analytical focus demanded for action at each level, is difficult for human managers with limited cognitive capacity. The prototype IISIS addresses this problem by organising information in the distributed knowledge base so that it may be accessed by jurisdictional level or by function across jurisdictional levels. The visual representation of information both via GIS maps and by graphic presentation through the IISIS screens facilitates this process for emergency managers, operating under the urgent constraints of time and limited resources characteristic of disaster environments.

The trial demonstration of the IISIS prototype sought to assess the extent to which the problem solving capacity of practicing managers may be enhanced through the use of a socio-technical decision support system. The simulated operations exercise provided a set of hypothetical problems to address the five issues identified by the emergency managers as problems in their current operating environment. Solutions for these problems were drawn from the information resources available through the prototype IISIS to demonstrate alternative strategies for action to protect life, property and continuity of operations for the community varying in time and cost.

Importantly, the practicing managers who observed the demonstration responded favourably to innovative uses of information technology in disaster management. In the evaluation survey distributed to managers who witnessed in the demonstration, 16 out of 16, or 100%, responded that the use of an information system such as IISIS would improve performance among agencies in their daily operations. Further, 17 out of 17 managers, again, 100%, reported they would be willing to invest their time and resources in learning more about the IISIS system. While these findings represent a small sample of local managers in the Pittsburgh Metro Region, they indicate that professional managers recognise the utility of a decision support system in emergency management and are willing to learn the skills and invest the time needed to integrate it into daily operations.


9 Conclusion

The IISIS prototype supports five critical functions that practicing disaster managers need to increase efficiency in disaster mitigation and response. These functions are:

1 Reliable, systematic exchange of information within and among organisations with legal responsibility for disaster operations and management.

2 Timely, accurate information to assess known threats to the community.

3 Full transfer of information from one jurisdictional level of operations to the next during disaster operations.

4 Real-time monitoring of threatening conditions, and the timely transfer of information from field units to operations centre and return.

5 Effective reporting of actions from multiple sites of operation, with the timely integration of these reports into a comprehensive profile of the entire disaster operations.

These conditions provide a structure for the timely exchange of information to practicing managers, but flexibility to adapt to emerging threats and changing conditions in an environment under threat. In our organisational planning process, practicing managers in the Pittsburgh Metro Region welcomed the use of information technology as a means of increasing their capacity to adapt their operations more effectively both to the demands of a changing disaster environment and to the actions of other organisations participating in response operations.

While there is no guarantee that a socio-technical decision support system such as the IISIS prototype will in fact solve the problem of coordination among multiple organisations engaged in a difficult set of operations such as emergency response, all indicators point to this effort as a promising strategy for creating the workable balance between order and flexibility required for coordination in practice. The results obtained from the trial demonstration project in the Pittsburgh Metro Region support this conclusion. The IISIS Project is continuing to develop more sophisticated models and further tests to validate this working balance in practice.

Acknowledgments

We acknowledge, with thanks and appreciation, past members of the IISIS Staff at the University of Pittsburgh who also contributed substantively to the design and implementation of the IISIS prototype in this research project. They are: Cheryl Delaney, Robert Blackwell, Liu Shuo and Jun Peng. We also acknowledge, with warm thanks, the contribution of my colleague and co-principal investigator for the IISIS Project during this trial demonstration, Dr. Hassan Karimi, School of Information Sciences, University of Pittsburgh. Further, we thank the practicing managers of the University of Pittsburgh, City of Pittsburgh, Municipality of Penn Hills, Wilkins Township, and the Southwestern Pennsylvania Chapter of the American Red Cross for their time and insights into the decision processes of disaster management. We are grateful to the Public Entity Risk Institute, Fairfax, Virginia, for its financial support of this research.


References Argyris, C. (1993) Knowledge for Action: A Guide to Overcoming Barriers to Organizational

Change, San Francisco: Jossey-Bass.

Axelrod, R.M. (1984) The Evolution of Cooperation, New York: Basic Books, Inc.

Axelrod, R.M. and Cohen, M.D. (1999) Harnessing Complexity: Organizational Implications of a Scientific Frontier, New York: The Free Press.

Caiden, N. and Wildavsky, A. (1974) Planning and Budgeting in Poor Countries, New York: Wiley.

Carley, K.M. (1998) ‘Organizational adaptation’, Annals of Operations Research, Vol. 75, pp.25–47.

Carley, K.M. and Lee, L. (1998) ‘Dynamic organizations: organizational adaptation in a changing environment’, Ch. 15 in J. Baum (Ed.) Advances in Strategic Management, Disciplinary Roots of Strategic Management Research, JAI Press, Vol. 15, pp.269–297.

Carley, K.M. and Harrald, J.R. (1997) ‘Organizational learning under fire: theory and practice’, American Behavioural Scientist, Vol. 40, No. 3, pp.310–332.

Chambers, R. (1974) Managing Rural Development: Ideas and Experience from East Africa, Uppsala: Scandinavian Institute of African Studies.

Chisholm, D. (1989) Coordination Without Hierarchy: Informal Structures in Multiorganizational Systems, Berkeley: University of California Press.

Coakes, E., Willis, D. and Clarke, S. (Eds.) (2002) Knowledge Management in the Sociotechnical World: The Graffiti Continues, London: New York, Springer.

Cohen, M.D. (1986) ‘Artificial intelligence and the dynamic performance of organizational designs’, in J.G. March and R. Weissinger-Baylon (Eds.) Ambiguity and Command: Organizational Perspectives on Military Decision Making, Marshfield: Pitman, pp.53–71.

Cohen, W.M. and Levinthal, D.A. (1990) ‘Absorptive capacity: a new perspective on learning and innovation’, Administrative Science Quarterly, Vol. 35, pp.128–152.

Comfort, L.K. (1989) ‘The logic of uncertainty’, Working Paper, 89–1, Berkeley: Institute of Governmental Studies, University of California, Berkeley.

Comfort, L.K. (1994) ‘Self organization in complex systems’, Journal of Public Administration Research and Theory, Vol. 4, No. 3, pp.393–410.

Comfort, L.K. (1999) Shared Risk: Complex Systems in Seismic Response, Oxford, New York: Pergamon Press.

Coovert, M.D. and Thompson, L.F. (2001) Computer Supported Cooperative Work: Issues and Implications for Workers, Organizations and Human Resource Management, Thousand Oaks, CA: Sage Publications.

Feldman, M.S. (1989) Order Without Design: Information Production and Policy Making, Stanford, CA: Stanford University Press.

Feldman, M.S. and March, J.G. (1981) ‘Information as signal and symbol’, Administrative Science Quarterly, June, Vol. 26, No. 2, pp.171–186.

Fesler, J.W. and Kettl, D.F. (1991) The Politics of the Administrative Process, Chatham, NJ: Chatham House Publishers, Inc.

Gell-Mann, M. (1994) ‘Complex adaptive systems’, in G.A. Cowan, D. Pines and D. Meltzer (Eds.) Complexity: Metaphors, Models, and Reality, MA: Addison-Wesley Publishing Co.

Hayes-Roth, F., Waterman, D. and Lenat, D. (1983) Building Expert Systems, Reading, MA: Addison-Wesley Publishing Company.

Holland, J. (1975) Adaptation in Natural and Artificial Systems, Ann Arbor, MI: University of Michigan Press.

Holland, J. (1995) Hidden Order: How Adaptation Builds Complexity, Reading, MA: Helix Books, Addison Wesley Publishing Co.


Hoschka, P. (Ed.) (1996) Computers as Assistants: A New Generation of Support Systems,

Mahwah, NJ: Lawrence Erlbaum Associates, Publishers.

Kaufman, H. (1960) The Forest Ranger: A Study in Administrative Behavior, Baltimore: Johns Hopkins Press.

Kauffman, S.A. (1993) The Origins of Order: Self-Organization and Selection in Evolution, New York: Oxford University Press.

Landau, M. (1969) ‘Redundancy, rationality, and the problem of duplication and overlap’, Public Administration Review, Vol. 29, pp.346–358.

Landau, M. (1991) ‘A multiorganizational systems in public administration’, Journal of Public Administration Research and Theory, Vol. 1, No. 1, pp.5–18.

La Porte, T.R. (Ed.) (1975) Organized Social Complexity, Princeton, NJ: Princeton University Press.

La Porte, T.R. and Consolini, P.M. (1991) ‘Working in practice but not in theory: theoretical challenges of high-reliability organizations’, Journal of Public Administration and Theory, Vol. 1, No. 1, pp.19–48.

Miller, G. (1967) ‘The magical number seven, plus or minus two: some limits on our capacity for processing information’, in Psychology of Communication, New York, NY: Basic Books, pp.14–44.

Ostrom, E. (1998) ‘A behavioural approach to the rational choice theory of collective action’, American Political Science Review, Vol. 92, No. 1, pp.1–22.

Peitgen, H., Jurgens, H. and Saupe, D. (1992) Chaos and Fractals: New Frontiers of Science, New York: Springer-Verlag.

Roberts, K.H. (Ed.) (1993) New Challenges to Understanding Organizations, New York: Macmillan.

Rochlin, G.I. (1993) ‘Defining high reliability organizations in practice: a taxonomic prolegomenon’, in K.H. Roberts (Ed.) New Challenges to Understanding Organizations, New York: Macmillan, pp.11–32.

Rosenau, J. (1997) Along the Domestic-Foreign Frontier: Exploring Governance in a Turbulent World, Cambridge, UK: Cambridge University Press.

Simon, H.A. (1981) The Sciences of the Artificial, 2nd edition, Cambridge, MA: MIT Press.

Simon, H.A. (2000) ‘Public administration in today’s world of organizations and markets’, PS: Political Science and Politics, December, Vol. XXXIII, No. 4, pp.749–756.

Taylor, F.W. (1967) The Principles of Scientific Management, 2nd edition, New York: W.W. Norton.

Waldo, D. (Ed.) (1971) Public Administration in a Time of Turbulence, Scranton, PA: Chandler Publishing Co.

Weick, K.E. (1993) ‘The collapse of sensemaking in organizations: the Mann Gulch disaster’, Administrative Science Quarterly, Vol. 22, No. 3, pp.606–39.

Wilson, J.Q. (1989) Bureaucracy: What Government Agencies Do and Why They Do It, New York: Basic Books.

Notes 1 Robert Axelrod and Michael Cohen, in their 1999 book, Harnessing Complexity, use the term

‘agents’ to mean any components that respond to information and interact in a systematic way. The agents may be individuals, organisations or computers, or combinations of same.

2 These needs are summarised from interviews with practicing disaster managers after local disasters in Pennsylvania, as well as 14 major earthquake disasters. Please see L. Comfort, J. Abrams, J. Camillus and E. Ricco 1989. ‘From crisis to community: the Pittsburgh oil spill’,


Industrial Crisis Quarterly, Vol. 3, No. 1, pp.17–39. See also L. Comfort and A. Cahill 1988. ‘Increasing problem solving capacity between organizations: the role of information in managing the 31 May 1985 tornado disaster in Western Pennsylvania.’ in Managing Disaster: Strategies and Policy Perspectives, Durham, NC: Duke University Press: pp.280–314. L. Comfort. 1999. Shared Risk: Complex Systems in Seismic Response. Oxford and New York: Pergamino Press. L. Comfort. 2000. ‘Response operations following the Chi-Chi (Taiwan) earthquake: mobilizing a rapidly evolving, interorganizational system’, Journal of the Chinese Institute of Engineers, Vol. 23, No. 4, pp.479–492. L. Comfort and Y. Sung. 2001, ‘Organizational learning in crisis: the 1999 Marmara and Duzce, Turkey earthquakes’, in U. Rosenthal, L. Comfort and A. Boin, (Eds.) 2001, Managing Crises: Threats, Dilemmas and Opportunities. Springfield, IL: Charles C. Thomas, Publisher: pp.119–142 [In press].

3 To identify the patterns of information search, exchange and feedback during emergency operations in the Pittsburgh Metro Region, a set of 27 interviews was conducted with practicing emergency managers from the six participating entities – University of Pittsburgh (8), City of Pittsburgh (5), Municipality of Penn Hills (5), Town of Wilkins (3), Allegheny County (5) and the Southwestern Pennsylvania Chapter of the US Red Cross (1) [8+5+5+3+5+1=27]. The interviews for the University of Pittsburgh were done as the first stage of this project in February–March, 1999. The interviews for the other five entities were done in May–July, 2001. To update the information for the University system, the eight managers were reinterviewed in September to confirm the initial map of information flow for the University and to add any changes that may have occurred since 1999.

4 For example, Allegheny County’s efforts to initiate a decision support system for emergency management in 1994 and 1995 were abruptly stopped by a change in the governing party in the 1996 County elections. The incoming administration, acting to keep a campaign pledge to limit the size of government in the County, cut personnel in County agencies by 20% and eliminated the County GIS Office altogether. Pittsburgh Post Gazette, January–March, 1997. Interview, Michael Mayhak, former Director, GIS Office, Allegheny County, January 24, 2001. That action set back the development of a viable GIS for Allegheny County that would be available for emergency management by at least four years. Fortunately, beginning in January, 2001, the effort to rebuild the GIS capacity for the County was renewed, and IISIS staff members worked collaboratively with County staff to rebuild the GIS capacity of the County for emergency management.

5 See the Incident Report page on the IISIS web page: http://www.iisis.pitt.edu.

http://www.jishin.ucsur.pitt.edu./

Coordination in complex systems: increasing efficiency in disaster mitigation and response

Documents

Transcript of Coordination in complex systems: increasing efficiency in disaster mitigation and response