Part I : Simple Assignments

This work was motivated by a talk I had with an employee of a home service company.

Assume you are the manager of several workers that must perform home visits. Your company has a large coverage area of about 60 miles in diameter. Your workers live as well as service homes within this area. The problem is, how to schedule workers to jobs sites to minimize their travel time. Ideally, your company would have a worker service the neighborhood he lives in. However, that is an unlikely scenario.

... (continue reading)

Part II : Multiple Worker Assignments

Continuing with the line of work prompted by a conversation I had with a home service company, we now consider the assignment problem for multiple workers.

Assume you are (still) the manager of several workers that must perform home visits. Your company has a coverage area of 60 miles in diameter. Your workers live, as well as service homes, within this area. Here, in Part II, we will address the problem of how workers should be scheduled to jobs to minimize their travel time.

... (continue reading)

Part III : Multiple Day, Multiple Worker Assignments

Further extending the work completed in Part II, we will add another dimension making the problem even harder.

Assume you have not found another job and you are still the manager of several workers that must perform home visits. Your company now has a coverage area of 61 miles in diameter. Your workers live, as well as service homes, within this area. Here in Part III, we will consider how to address the problem of assigning workers to jobs for multiple days to minimize their travel time.

... (continue reading)

Part I : The Effect of Demand Fluctuations

Investigating how commercial supply chains react to drastic swings in demand. This is a very generalized model to show how the inventories of a multilevel supply chain, ex. manufacturer, warehouse, retailer, and customer are affected.

Part II : The Effect of Illness and Interactivity

The effect of illness on a business' labor force is investigated, with implications of contact rates related to staffing and general operations decisions. Some of the precautions to minimize infections rates have been implemented in our society.

Part III : Modelling A Local Humanitarian Supply Chain

The HSC focuses on customer demand fulfillment (instead of profit) and can also be severely impacted by drastic demand/supply fluctuations. Assumptions were made about the general operations of a HSC service provider. A framework for managing the suppliers operations was developed.

A quick introduction to Power Automate. It provides hundreds of alerts/actions to use to build a custom work flow. In this article, I combine a simple form with two inputs with some Twitter and LinkedIn actions.

Part I : Roads and Cars

With the hypothesis that traffic flow benefits from having both slower and faster moving individuals through traffic, a model was developed to test it. First, the types of entities, roads, and cares, and their actions are defined.

Part II : Effect of Traffic Velocity Variance

Continuing, the effect of the velocity variance of the population of cars is determined. Specifically, comparing the scenario where the range of velocities is approximately between 30 and 40 mph and the scenario with velocities within the approximate range of 15 and 55 mph.

Part III : Effect of Traffic Density and Disabling Lane Changes

Two other experiments were run to investigate 1.) the interaction of the velocity variance and (dis)enabling lane changes and 2.) velocity variance and the rate of new car arrivals. The two experiments are independent (and could not be directly compared). The experimental results did provide some insight into what driving behaviors are not beneficial to traffic flow.

Exploring Google's Kaggle.com, a home for data scientist and their data sets, a data set for natural language processing (NLP) was found. It contained titles and text of more than 6,000 articles. Here, these articles were identified as either real or fake news.

Part I : Introduction

This article starts with an introduction to simulation, specifically discrete event simulation (DES). The modeled system is a sandwich restaurant that has a order taker, two cooks and a limited seating area. The arrival rates of customers and worker processing times are defined here.

Part II : Implementation with Simpy

After defining the system, in this article, using the free and open source DES package, Simpy, the model is formalized. The methods for customer arrivals/balks, order preparation, and dine-in or carry-out choices are defined. The results of a simulated day are then interpreted.

Part III : Implementation with Simio - Personal Edition

Using a Personal Edition (free but limited) version of Simio, the system presented in Part I is modeled. The same process of defining customer arrivals/balks, order preparation, and dine-in/carry-out options is done. The results identify potential areas of improvement.

Using SciPy for Linear Math Models

This article introduces a small portion of the SciPy.Optimization package, a linear programming solver, to solve the product mix problem. This problem is the need to determine how much and which products to make given that all share input resources and require different constraints to build.

Using Google's OR-Tools To Solve The Shortest Path Problem

Google's OR-Tools have several solvers for generalized problems. Here, one of which, the shortest path problem, is formulated and formatted to be solved in one of the OR-Tools.

Using COIN-OR Symphony To Solve a Scheduling Problem

COIN-OR has existed for some time. In this article, the shift scheduling problem, as was presented in Winston, is solved using the Symphony package.

Part I : Crime Time and the Times of Crime

The impact of time on crime, specifically day of the year, was investigated. Before reading, take a moment to guess which days may have had the highest and lowest number of occurrences. Also, based on the number of occurrences, best fitting trends were calculated that could potentially be used in forecasting the number of occurrences in successive years.

Part II : The Crimes at Crime Time - Not Time Crimes

This article will investigate the types of offenses that are committed, with the later sections focusing on the identified relatively higher number of offenses days identified in Part I. Before reading, take a moment to guess what types of offenses may have occurred most frequently throughout the year, and more specifically, on the high number of offense days.

Part III : The "@" of Crimes at Crime Time

Included in the APD data set is latitude-longitude coordinates. The locations of the crimes, crossed with the previous analyses show some interesting interactions. In this last article the location of the incidents for specifically the two highest crime days was evaluated.

The latest release of data collected in 2016-17 which queried survey participants about their unhealthy behaviors, preventative actions, and possible health outcomes was analyzed. In this article, some exploratory data analysis was done to reveal some basic information and then extrapolated to determine what other inferences can be derived from the data.

In this article an image categorizer was developed to investigate the internals of a neural network. The process of manipulating an image as it passed through the network, and ultimately, how the data is manipulated to get an accurate prediction was visualized.