NPHS 1510: Federal and International Framework
Emergency Problem Solving, Decision Making and Risk

 
Competitive Decision Situations 
 
There are a number of decision situations in emergencies where the decision maker/emergency manager is working against one or more other agents whose actions can cause the decisions to be erroneous or sub-optimal. Some examples are:
 
  • The behavior of humans or animals in an emergency situation or in reaction to a decision
  • Criminal Behavior
  • Terrorist Behavior
  • The behavior of multiple units or jurisdictions in preparing for responding to or recovering from an emergency
The diagram below shows a schematic of the decision process in competitive situations.
The actions of the competitors can be sequential, coming after the manager's decision or simultaneous, occuring in the same time frame as the manager's decision. Sequential decision situations are instances where later actors may have some knowledge about earlier actions. These competitive situations are quite similar to games and often game paradigms are used to teach the decision principles. The study of competitive decision situations is called game theory.
 
Definition     Game Theory: the study of the ways in which strategic interactions among economic agents produce outcomes with respect to the preferences (or utilities) of those agents, where the outcomes in question might have been intended by none of the agents.

Source: Stanford Encyclopedia of Philosophy.
 
Some classification of competitive decision situations are:
  • Collaborative vs Non-Collaborative - participants can work together to achieve a goal or work against each other.
  • Single Participant vs Multiple Participant - participants can work individually or as teams.
  • Discrete vs Continuous - discrete games have a finite number of actors, events, actions, outcomes, etc.
  • Symmetric vs Asymmetric - symmetric decision situations are ones where the actors select from the same strategy set and have the same outcome values.
  • Zero-sum vs Non-zero-sum - In zero-sum decision situations, the total outcome value to all participants, for every combination of strategies, always adds to zero (i.e. when one actor gains, one or more of the other participants lose a total equal to the gain of the actor).
 
Perhaps the most common competitive decision situation that we have all encountered is the game of Tic-Tac-Toe. Tic-Tac-Toe is a sequential, non-collaborative, two-player, symmetric, discrete zero-sum game. The elements of Tic-Tac-Toe are:
  • Players - 2, each player is assigned a mark: X or O
  • Arena - A 3x3 grid of cells
  • Start State - All of the cells are empty
  • Actions - Place a mark in a cell
  • Rules - X has the first move, moves alternate between players, a mark can be placed only in any open cell
        
  • Goal - Be the first player to arrange three of their marks in a straight line, otherwise the game is a tie
    Example:    
Tic-Tac-Toe is a game of perfect and complete information. Each player can know the structure of the game (strategies and payoffs of the other player) and the moves previously made by the other player.
 
In most real competitive decision situations, the state of knowledge of the decision maker is far from perfect or complete. This is particularly true in emergency situations. For example, the emergency manager may or may not know:
  • Who the other actors are.
  • What their goals are.
  • What outcomes might result.
  • What strategies are available to the other actors.
  • How the other actors evaluate the strategies and outcomes.
  • Which strategy is selected.
Acquiring the answers to these and other questions is what situational awareness is all about.
 
The following sections provide some brief examples of emergency competitive decisions and how they might be approached. The use of games for teaching decision making and improving the decision makers awareness and sensitivity are discussed. For a more intensive examination of the subject, consult the references listed at the end of this session.
 
The Arsonist Game
An arsonist has been burning down certain homes in a residential neighborhood. The map below shows an aerial view of the neighborhood with one of the destroyed houses marked with a red dot. Your goal is to discover the characteristic of homes that the arsonist is targeting as quickly as possible so that you can catch him or her.

Use your mouse to click on any house in the area. If that house is a target house, the house will be covered by a red dot. Otherwise it will be covered by a green dot. The game will keep track of the number of target and non target homes you have clicked.

The game ends when you have discovered all of the target homes. A better score will have a higher ratio of targt to non-target clicks (i.e. you have discovered the arsonist's pattern quicker). The link below the game gives the arsonist's strategy.

 
Click here for an explanation of the arsonist's strategy.
 
Search and Rescue Game
 
Computer video simulations have become another tool for learning how to handle competitive decision situations. The images below show screen shots from a game for search and rescue. In the game, a player has to develop strategies for locating objects/victims in a building or other environment. The search environment might be a replica of actual buildings that the student is familiar with or completely new locations. The game features single player and team modes. In team mode, the players learn to collaborate on searches so that they find the goal object quickly.
 
    
 
The simulation has a number of different search envionments: multistory buildings, sprawling buildings, factories outdoor environments, forests, etc. The objects of the searches can be placed anywhere within the environment. Some objects such as perpetrators can move to elude the search. Computer based training environments have become very popular because they are relatively inexpensive, can represent a wide variety of situations and can provide a near-real experience.
 
Terrorism Example
Often it is useful to recast the competitive decision situation to multiple decisions under uncertainty. Cosider the case of a terrorist planning an attack. The emergency manager must prevent or minimize the attack.

One aproach is for the manager to put him or herself into the shoes of the terrorist. The manager views the targeting decision as an uncertain problem. From the terrorist's viewpoint, there are a number of candidate targets, each with damage levels and likelihoods of success.

The tables below contain the decision matrices for a terrorist for a small town. These tables are the town emergency manager's assessment of:
  • The targets that the terrorist would hit
  • The likelihood that the terrorists would be successful for different levels of attack
  • The estimates of costs of outcomes
Note that this is the emergency manager's estimate of how the terrorists would view the situation.

 Target: Town Hall 
   Destroy  Damage   
 Probability  .7  .3   
 Outcomes  $2M  $.5M  1.55M 
      
 Target: Power Plant 
   Destroy  Damage   
 Probability  .2  .8   
 Outcomes  $3M  $1M  $1.4M 
      
 Target: National Fabric Museum 
   Destroy  Damage   
 Probability  .6  .4   
 Outcomes  $4M  $1M  $2.4M 
 
In this case the emergency manager recognizes that the National Fabric Museum is the optimal target from the terrorists point of view (i.e. where the terriorist can do the most damage).

The emegency manager has a number of prevention and mitigation strategies available to address the terrorist problem. These strategies may be general or asset specific. They include but are not limited to:
  • Increases police patrols.
  • Intelligence gathering (e.g. cameras, informants, etc.).
  • Traffic rerouting.
  • Awareness campaigns.
  • Access restrictions (IDs, permits, etc.).
  • Inspections (Scanning, searches, etc.)
Any action that the emergency manager takes will affect the the decision problem of the terrorist. The end result, may be a total change of how the terrorist perceives the decision environment (i.e. his or her situational awareness).
 
Suppose the emergency manager decides that the real threat lies at the National Fabric Museum. The manager also decides to use a deterrence strategy that involves installing a vehicle barrier system similar to the one shown below.
 
 
Installing the vehicle barrier would likely cause a terrorist to reassess both the probability of success of an attack as well as the damage that the would occur. The new decision matrix for the National Fabric Museum attack is:
 
 Target: National Fabric Museum 
   Destroy  Damage   
 Probability  .4  .6   
 Outcomes  $3M  $.5M  $1.5M 
 
Several observations and conclusions are worthy of note:
  • The expected damages from an attack on the museum have been reduced. The emergency manager must evaluate whether real cost of installing the barrier is worth the reduction of expected cost of damages.
  • From the terrorist's viewpoint the problem has changed. Attacking the museum is no longer the desirable strategy. Attacking City Hall is.
  • What our actions have done is to restructure the problem. The emergency manager now has to deal with the prospect of an attack on City Hall.
The situation is not unlike the Cold War where the introduction of a countermeasure evokes a new threat or weapon. The emergeency managers ultimate goal is to find a strategy or set of strategies that cause the terrorists to abandon their plans.
 
Exercise     Outline a decision situation for a hostage standoff as a game. Identify the actors, strategies, outcomes, decision rules, etc. Demonstrrate how your game might play out for a specific scenario.
 
Resources     Game Theory.net

Game Theory Society
 
              

Copyright © 2011 Ken Sochats