Public Systems Modeling:
Systematic identification and evaluation of public policy alternatives.
We use mathematical models to do this.
Two main types of mathematical models:
Optimization: Finding the best of many alternatives.
Addressing ‘What should be’ questions.)
Simulation:
Finding the impacts of alternative assumptions
(Addressing ‘What if’ questions.)
Mathematical models:
Consists of decision variables (unknowns) and parameters in equations and inequalities that together define a system.
Example – design a cylindrical tank to hold a given volume of maple syrup:
Define objective that is used to rank or evaluate alternatives.
Define the constraints that must apply.
Objective: Minimize Tank cost subject to Volume requirement (constraint).
Identify data needed:
Cost per unit area of surface of tank. storage capacity (volume) required
Construct model: (Make sure units of each term are consistent.)
Min: [(Cb+Ct) ($/area)]*[(π*r2) (area)] +
Cs ($/area) ( 2 π r h) (area) = $ π*r2h ≥ Volume
We use procedures (algorithms) to solve this and other models.
Types of Models.
•

Conceptual models - cause and affect diagrams – often a precursor to more quantitative mathematical models.

•

Simulation = What if? Descriptive models.

•

Search
Systematic iteration, Gradient hill climbing, Evolutionary, Heuristic.

•

Decision analysis – node-link networks where links are possible decisions or chance events, nodes are states of the system where decisions are to be made or where events will happen.

•

Optimization = What should be? Prescriptive models.
Deterministic and Probabilistic or Stochastic
Linear and Non-linear
Dynamic and Static etc…

•

Decision support systems – shared vision modeling

Why Model? To identify and evaluate alternatives.
•
•
•

Estimating their impacts (econ, social, environmental and ecological, financial, etc.)
Assist decision making by providing useful information,
Identifying and presenting tradeoffs among conflicting objectives.

Decisions will be based on more than model output.
On qualitative and quantitative information.
“All models are incomplete and wrong.” The real world is more complex.
Models are often simplifications of reality.
Under what conditions might modeling be useful?
•
•
•
•

When some decision is to be made regarding some issue or opportunity.
When there are many alternative decisions that could be made.
When the best alternative or