Date of Award

December 2015

Degree Type


Degree Name

Master of Science



First Advisor

Wilkistar Otieno

Committee Members

Hamid Seifodinni, Todd R. Miller


Active Pharmaceutical Ingredients, Agent Based Modeling, Simulation Modeling, Wastewater Pollutants


Pharmaceutical pollutants are present in traceable concentrations in Milwaukee County water system and Lake Michigan. The actual point sources and nature of entry into the water system is difficult to determine with certainty. Pharmaceuticals have been found to persist at the South Shore Wastewater Treatment Facility (SSWTF) in Milwaukee, Wisconsin. The highest concentration was found to be for the pharmaceutical drug metformin. Metformin is a first line drug for the treatment of type 2 diabetes mellitus.

The broad goal of this exploratory study; the first of its kind, is to correlate trace concentrations of drugs to the point sources. Particularly, we have analyzed the demographic and geographical location of the population in Milwaukee County as potential contributors of drugs in the waste water system. The study uses metformin as a pilot study. The objectives of the Thesis are as follows:

Objective 1: Analyze the current status of quantifiable pharmaceuticals in Milwaukee County water

Objective 2: Analyze Milwaukee County’s geographic, demographic and socio-economic status to determine and quantify the main population agent attributes to be used in objective 3.

Objective 3: Build an Agent Based Simulation Model that incorporates county residents as API point source agents, whose attributes as listed in objective 2 to determine their degree of contribution towards APIs in the waterways.

We were able to develop a working model and verify and validate the results through publicly available empirical studies and records. No studies were found utilizing simulation modeling to determine specific sources of pharmaceutical drugs entering the water system. No studies were found measuring the effect of age, race, income and geographical location of potential drug contributors on pharmaceuticals found in the water system. In this study, the results for the total number of diabetics in Milwaukee County was determined to be within an acceptable margin of error specified in the model. The model results indicate approximately 91,353 type 2 diabetics in Milwaukee County which is corroborated by Wisconsin’s department of health reports indicating 93,020 hence the results are within 2% margin of error. The adjusted contributors that account for the measured concentration of metformin at the SSWTF were determined to be within the range of concentrations measured in the influent stream – Minimum: 3,200ng/L, Median 55,000ng/L, Maximum 100,000ng/L. The model results indicate that 76,899.3ng/L concentration present from the contributors which is within range of the actual measured concentration of metformin at the SSWTF influent stream. The model output includes geographical, income, race and age demographics of the contributors by count for each of the 34 zip codes in Milwaukee County. Results analysis indicate that a geographical, racial and socioeconomic difference exist in contributors overall for Milwaukee County. The results also confirm the prevalence of the drug metformin entering the water system while identifying a zip code level detail of individual contributors. This Thesis is an exploratory test bed example showing that agent based modeling can be a valuable tool for industrial engineers and operation research when dealing with geospatial problems that exhibit variability through agency.