Hwk #1

Liebman, Jon (1976) “Some simple-minded observations on the role of optimization in public systems decision making,” Interfaces 6(4) pp. 102-108.

Summary:
At the time when this article was written, optimization models using linear programming to calculate the best solution to a problem were being widely used in a number of fields—notably in the public sector and in improving firefighting strategies. In firefighting, models had been created to calculate where fire trucks and fire stations should be placed, and these models had been shown to increase the effectiveness of firefighting operations. In the public sector, particularly river basin quality management, the author could only find one example where the optimization models had been used with any success.
According to the author, those problems which had successfully been solved using optimization models had several common traits: they were problems of increasing efficiency, the goals of the model as well as its constraints were obvious, and since it was the private sector making the decisions, there was only one stakeholder. As the linear programming methods have become more complicated, the problems these models are used to solve have become more complicated, particularly those tied to the public sector. Liebman describes these problems as “wicked problems.” Common characteristics of the wicked problems include highly interconnected systems with a very large number of stakeholders, as well as complicated problems where the results of certain actions are unknown.
Liebman says that when solving these complicated models, the role of the model changes. Instead of calculating the most efficient solution, models are now used to calculate a number of different solutions whose purpose is to aid the decision-maker in maker the final decision.
Discussion:
Liebman showed how the methods to using models undergo profound shifts when dealing with complicated problems versus with simplistic problems. His four suggestions contained near the end of the paper were interesting, but I found his first two to be the most insightful; the gist of these two suggestions being that a complicated model is actually the organized thinking process of an individual, and therefore there are many possible models for a single problem.
I feel like further research along this same line could be to show what type of models the different stakeholders of a public problem will develop. Since, according to Liebman, more models help the decision-maker, having models representing a wide array of perspectives in addition to scientists and engineers is vital.



Ostfield, Avi and Salomons, Elad. (2004). "Optimal Layout of Early Warning Detection Stations for Water Distribution Systems Security.” Journal of Water Resources Planning and Management. 130(5), 377-385.

Summary:
Since September 11, the danger posed by evildoers to public water distribution systems has been a point of concern to the EPA and public utilities. The EPA has been funding increased security for water distribution systems, promoting information sharing between the various institutions, and encouraging improvements to the detection and treatment methods used by the local water utilities. The ability to monitor water quality within a distribution system is of particular concern, since early warning of contamination can provide valuable time for the utility to implement life-saving counter-actions.

In their paper, Ostfield and Salomons attempt to improve the way the early warning systems are laid out. In a best-case scenario, water quality would be monitored at every node in a distribution; however available technology makes this cost-prohibitive. In conjunction with the monitors, chlorine boosters are placed in a system, and the optimal locations of these boosters may be calculated using linear modeling. There are several linear programming models in existence create a binary matrix which calculate contamination at each node over time, calculating each node as a potential source. These models have several shortcomings: they only consider steady state conditions and they do not consider residence time. Also, since these models assume that water upstream of an acceptable node will be acceptable, they encourage placing monitors on the edges of the distribution system, which means that contaminants within the system may not be detected as soon.

Building off of these methods, Ostfield and Salomons developed a linear program model which can calculate pollution at nodes using a similar method to the old approach, except that their model allows simultaneous contamination from multiple nodes, and their model uses a complicated algorithm to calculate the evolution of biological contaminants. Using their model for two simulation water distribution systems, the researchers were able to calculate the level of service versus the number of monitoring stations.

Discussion:
I felt that the method used by the researchers could provide some interesting insight into monitoring water quality. I particularly thought that their including multiple contamination sources as well an algorithm for the movement and evolution of biological contaminants could be useful.

I feel that their research, while it may be useful in finding an optimal number of monitoring stations, does not address the problem of placement of these stations. Once the number of monitoring stations has been found using the methods detailed in this paper, further research could be done to develop a method for the optimal placement of the stations throughout the system.