Week 7 : Research and Debuging thoughts

Hi guys, what’s up? In the past week I kept on debuging the code while researching on the pros and cons of computer simulations for my final research article. Here are some of my thoughts:

There are always limitations and drawbacks of genetic simulations comparing to empirical data in representing scenarios in real life. In order reach the scale needed to have statistical significance similar with empirical data, many loops, recursions, and random generation(creation) of data need to be accounted for, creating lots of possibilities for errors, or “bugs”, to occur. Although this is unavoidable for all long and complex programs, the structure of simulations means the functions are highly interconnected and indexing is difficult, which makes the finding and solving of these problems very hard. Worse, bugs like index errors don’t affect the outputting of the codes, therefore the user might unknowingly receive a completely false result if there aren’t empirical data to compare to. ………

In addition to high amount of error-prone data, assumptions also must be made, which could make the simulations less realistic. For example, in our simulation, we are assuming a perfect square grid field that the plants grow on. Although different survival rates are assigned to center and edges of the grid, there are still many independent variables that we have to assume don’t affect the survival or reproduction pattern of the plants: slope of the field, soil fertility on different parts of it, and interaction with other species that it compete with space and resource for. All of these variables hinder the accuracy of the simulation in representing scenarios in real life.   …… Of course, given unlimited effort a programmer could set a function to account for each of these conditions, but that would lengthen the code exponentially and thus defeat the purpose of accessible and simple experiments—a major strength of simulations versus empirical studies.