Smart city take home question answers

1. Take Home Final Exam/Project
Aydin Ayanzadeh
Department of Computer Engineering
Email: ayanzadeh17@itu.edu.tr
Scenario 1
Waste collecting process in scenario 1 is not very complicate. Actually, we have truck with that start
its mission and traverse each trash-bin every hour and collects the waste from them(we assume
that, the truck has hypothetical speed and end up its mission during the one hour). If we want to
simulate this problem to the algorithms or methods, it is easily predicted that, this is similar to
TSP(Travels salesman problem) that truck traverse all vertices and come back to its source. But
TSP algorithm is NP algorithm and has high complexity in during the execution with many number
of nodes, therefore i have to implement the code with greedy, approximation and heuristic method,
scenario 1 is the general model of of the scenario 2 that x=18. in scenario 1, i implement this
scenario by Dijkstra method (for calculating the shortest path between the S and D) and also, i
specialized and D to s). Moreover, The path (shortest path) from S to D is 24 and ta backward path
to the trash center based on dijkstra algorithm (dijkstra.m[1]) is 7 (red line in Fig1), P=240+70=310
Km, so the cost of truck in 24 hours with suggested path is cost=24*3.5*310=26040$ (minimum
cost for this path). Furthermore, this problem is not faced with overflow capacity of trash bin during
24 hours. I provide an greedy method besides the past method, number of edges is step and the
and height of rectangle is h=4, if rows of trash is odd the formula for this 24*3.5*320= 26880$.
Fig1,2. Result of scenario the left image represent the waste of trash bin in 21th hours of a day.
Right image shows the total cost of waste collection for 24 hours.
Scenario 2
In scenario 2, the problem is specialized in numbers of trash-bin, for solving this problem, I
use dijkstra algorithm for (n+1) times n=18 and store them in an (n+1)(n+1) matrix, each
rows of matrix represents the shortest path from vertex to all other vertices. I use the
BLG556E-Digital Solutions for Smart Cities, Spring 2017

2. shortest near neighbors in this problem, i start the path from trash collection center and find
the shortest path to all vertices and find them nearest one. Afterward, connect the neighbor
edges to the nearest one and keep this way. Nearest method shows the shortest path for
different x =2,4,6,8,10 (my method can not evaluate the cost of graph for different X,
because it has faced with syntax error that it could not solved, but it can shows the nearest
neighbor for each edges)
fig3.The architecture of trash bins and trash center fig4.The simple shape of problem in graph mode
red line shows the shortest backward way for
b) Discuss what will happen when the capacity of a truck is fixed. Write your solution suggestions.
If we assume that they have infinite waste collection capacity, we will not encounter the problem in
waste collection process. Conversely, when the truck has fixed capacity, the probability of lack of
capacity is high in this condition. fix capacity in the municipality trucks leads to use efficient methods
of waste collecting of trash-bin. The common solution to solve this problem is increasing the
number of truck for each hour. But, scheduling and planning technique is more proper way to solve
this problem in fix capacity condition. The more efficient way to solve this problem is using priority
method for each trash bin based on their waste collecting every hour and scheduling for eachλ
trash bin based on its . By this technique, it is possible that in each iteration of collecting theλ
amount of waste is reduced to half of normal condition based on scenario 2.
C) Discuss how the trash collection method explained in Scenario#2 could be enhanced.
BLG556E-Digital Solutions for Smart Cities, Spring 2017

3. The most important and significant issues in waste collection problem is avoiding from overflow of
the trash bins due to its additional financial loss to the municipality. For increase the efficiency of
waste collection, we should give priority for each trash-bin ​poisson random variable ​based on
its ( ). For example, the trash bin with trash_id (4.3) has =25 has more speed of increment ratherλ λ
than the speed of trash bin id 3.1 with =5; assume that trash bin with low is filled half of itsλ λ
capacity, in this condition the sensor send an alarm to the trash center, conversely, if the trash bin
with higher has the waste between 20-25, in this condition, sensor does not send any alarm to theλ
center despite of its high probability to have overflow in next hour and this deficiency in given
system brings extra financial loss for trash center. Hence, by reducing the overflow trash bin and
number of trash bin that send alarm to the center, related organization will reduce its addition cost
of waste collecting.
Reference
[1] ​mathwork File exchange official websites.By Kashif Shahzad,3rd June 2004(dijkstra algorithm).
BLG556E-Digital Solutions for Smart Cities, Spring 2017