IISE Simulation Competition
In a group of three, we competed against school across Canada and the northeastern United States in a simulation competition and were awarded third place. The project focused on maximizing the efficiency of a construction project using Arena Simulation Software. We each shared roles and collaborated, although I focused more on the implementation and my teammates on the heuristics.
3rd Place Award
Group of 3
Tools Used: Excel, Arena, AutoCAD, Heuristics
Simulation Competition Presentation
In a team of three, we competed against schools across Canada and the northeastern United States in a simulation competition run by a student chapter of the Institute of Industrial and Systems Engineers and SimWell. These simulation competitions look at a complex process and the range of unpredictability in subprocesses and use the Arena Simulation Software to maximize efficiency. Given a problem, we create a solution, write a report, and submit it. Since we were selected as finalists, we then prepared a slideshow and presented in front of a panel of judges.
As a team, we were presented third place!
Reception of Simulation Award
Simulation Award Envelope
As part of a firm hired to optimize the construction process of a four-sector, four-level tower, we essentially used Excel to organize our data and then used Arena to implement our leading heuristics and simulate the randomness involved in the operation of the cranes.
Construction Site Layout
We were asked to provide a full schedule of what order the parts would be delivered in, with respective batches, an installation schedule, the position of each crane at any particular time, and budgets.
Objective and Deliverables
We had the option of using any combination of large or small cranes for any periods of time in any of four different placements. After this, we had control over which cranes picked up which parts and in what order. There was a set of staging areas that each of the parts could be placed in, or they could be directly loaded from a truck or two if a permit is purchased.
In brief, we determined an order of part arrival using Excel and a VBA macro, determined which cranes would be utilized with heuristics, implemented the solution in Arena, and analyzed the results. In the first portion of this, we simplified the problem as we realized the sets of parts between sections are the same, then used a macro to batch all of the parts such that the maximum capacity of each truck was utilized. Then in AutoCAD we recreated the job site and explored the different possible crane configurations.
All Possible Crane Layouts
We found that small cranes can only offload trucks as their reach does not allow them to place parts in any particular section. A preference matrix was then used to evaluate each configuration and select the best one given the covered surface area, the overlap of cranes, and cost. The strategy selected uses two small cranes to store parts and the opposite two cranes to build, each with limited degrees of freedom to prevent collisions while ensuring the building cranes were at maximum utilization.
Final Crane Layout
We then implemented the solution in Arena, and visualized the results with charts, animations, and metrics. This image is dynamic in the sense that the clock and date stamp update as time passes, and the yellows cranes change colour to green and the trucks bucket fills when they are in use.
Arena Visualization Screenshot
To further analyze the legitimacy of the model and the results, we graphed the utilization of each crane. In essence, the building cranes (2 and 4) are nearly always in use, while the storage cranes (1 and 3) have to wait for the stocking area to be cleared before it can place more parts, hence their much lower utilization.
By using the metrics tracked during the simulation and the batching results, we developed a set of schedules that resulted in 3 full days of labour (3 shifts per day), with 48 truck shipments and 2 large and 2 small cranes used for building and loading respectively. This resulted in a total of $268,360 and given our report and paper, we were awarded 3rd place!
As was seen in the crane utilization charts, the small cranes were idle plenty of the time. This was mainly due to the idea that they were only used for stocking as they could not reach all areas of their nearest section. Considering these cranes were able to reach some parts of those sections they could potentially build instead of stay idle. It could be worthwhile exploring this possibility; although, it would add a lot of complexity.