liu-hy / food-bank-network-simulation Goto Github PK

@Liu-Hy

Food class' subtraction function doesn't seem to account for the case when an order is greater than the available quantity.

Is this the intended behavior?

I should say that I would prefer it to still perform the subtraction (even if we end up with negative quantities) and the negative be handled outside the subtract func.

Related to #20 and #1

related to #1

I will assume the food bank purchases the freshest food always using the max_days global. Related to #1 and #7.

What do you think @Liu-Hy @lucianli123

To simulate increased load, it will be necessary to increase households per pantry and food insecure population on FoodBank init

Would be useful to have a Global class with our constants and variables that need to be accessed throught the project (like days and food price).

Motivation: better type hinting than dict; organization

Is there already a way for the food bank to give the pantry the food asked?

@Liu-Hy : Line 361 of FoodPantry.py returns None if not sunday

@rodigu : Line 46 of FoodBank.py doesn't properly handle None return from pantry.run_one_day

TypeError: cannot unpack non-iterable NoneType object

Still need to implement 3 methods

• hold a pantry event
• predict the demand in the upcoming event
• calculate utility

The operation of food pantries should be the main computation bottleneck. We can try multiprocessing. My idea is to load the foodbanks into several processes in the simulation file. The global coordination of purchase may be done after that in the main process.

Can @Liu-Hy add a get_demant function that returns the demand for each of the products pls

It can return a dict with keys being the types and values being the demand in weight.
I need the real demand, regardless if ir was met or not. This is to calculate how much of each food the bank will need to buy.

Should we have the pantry save some of the budget it receives?

Or can I spend it entirely on food?

Adding decorators in Numba to speed up our program is not as straightforward as expected. I met some problems.
We can try with different methods for selective compilation.
https://numba.readthedocs.io/en/stable/user/index.html
https://canvas.illinois.edu/courses/22388/pages/unit-12-page?module_item_id=2075138

Move from Food Pantry to food bank, make storage variable actually

Suppose there are $n$ people and the demand $X$ conforms to the same Beta distribution, with the expectation $\frac{a}{a+b}$. The supply is only $nm$, where $0 \lt m \lt \frac{a}{a+b}$. Set a limit $k$ for everyone such that the expectation of $\operatorname {min}(X, k)$ is exactly $m$.
It does not require much precision. Are there any numeric methods to solve the integral equation, or should I simply look up a table using the monotonicity of the expectation with $k$? For either way, are there any useful toolboxes in R/Matlab/Mathematica?

@rodigu, can you implement a main function of foodbank plz? No need to buy or share food, by calling run_one_day(), it receives newly donated food, lets pantries run sequentially, and gets aggregate waste and utility. I hope to have some result on the pantries of a bank before presentation.