Download as pdf or txt
Download as pdf or txt
You are on page 1of 2

lOMoARcPSD|22101168

SIMULATION LP-GRAPHING
Step 1: Make a lookup table; where the headings are: Step 1: Identify decision variables, objective function and
outcome, lower bound and upper bound. constraints
Step 2: Graph constraints and identify feasible region with
Step 2: Determine the distribution & solve accordingly.
labelled corner points.
Uniform Distribution Exponential Distribution Step 3: Using the corner point method, make a table
● x=a+(b-a)*RN ● -(1/λ)*ln(RN) Step 4: Determine the possible LP outcomes:
● Note: λ=1/μ → Unique Optimal Solution Alternate Optimal Solutions
μ=1/λ
Step 3: Make a queueing simulation table (if needed)

Note: Although the feasible region is


unbounded, it is alternate bc there are
two minimum optimal solutions

Unbounded Problem Infeasible Problem

COMPLEX LP FORMULATION
Transportation Transshipment
● ① xi,j= # of units shipped from i to j ● ① xi,j= # of units shipped from origin i to intermediate j
yi,j= # of units shipped from intermediate i to customer j
ci,j= shipping cost per unit from i to j
si= supply/capacity in units at origin i si= supply/capacity in units at origin i
dj= demand in units at j dj= demand in units at customer j
vk= volume in units at intermediate k

● ③ (S≥D) → ΣΣxi,j ≤ si ● ③(S≥D) → ΣΣxi,j ≤ si


ΣΣxi,j = di ΣΣxi,j ≤ vi
xi,j ≥ 0 ΣΣyi,j = di
(S<D) → ΣΣxi,j =si (S<D) → ΣΣxi,j =si
ΣΣxi,j ≤ di ΣΣxi,j = vi
xi,j ≥ 0 ΣΣyi,j ≤ di
xi,j ≥ 0
yi,j ≥ 0

ΣΣxi,j- ΣΣyi,j= 0 (amt going into intermediate node is


what comes out of it)

Downloaded by nidzo balling (nebokaca@lyft.live)


lOMoARcPSD|22101168

Scheduling Multi-period production planning


● ① Fi= # of full time employees ● ① ri= regular production on day i
starting at shift i oi= overtime production on day i
Pi= # of part time employees ii= extra carried over as inventory on day i
starting at shift i
cr= regular production capacity
● ② MIN: z= co= overtime production capacity
($x/hr)(#hr)(ΣFi)+($y/hr)(#hr)(ΣPi) di= production demand on day i

● ③ Full time work 8 consecutive hrs, ● ③ ri ≤ cr


part time=4: oi ≤ co
Fi + Fi+1 + Pi+1 ≥ c
Suppose that object produced in a day can be used to either
2 Full time for 1 part time: meet that day’s demand or be held for later use. No inventory
Fi ≥ 2Pi should be left on last day.
Day 1→ ri + oi - ii = di
Day 2-4 → ri + oi + ii-1 - ii = di
Day 5 → ri + oi +ii-1 = di

● BEG Inventory + Production - END Inventory = Demand

INTEGER PROGRAMMING
Integer Program Logical Constraints with Binary Variables

Let xA = 1, if project i is done, where


= 0, otherwise

“Project A must be done” xA =1

“Exactly 2 of projects A,B,C must be done” xA+ xB + xC = 2

“At most 2 of projects A,B,C must be done” xA+ xB + xC ≤ 2

“At least 2 of projects A,B,C must be done” xA+ xB + xC ≥ 2

“If project A is done, then B must be done” xA = xB


Step 1: Find the optimal linear solution
“If project A is done, then B must NOT be done” xA+ xB ≤ 1
Step 2: Using the integer feasible
solutions, the optimal integer solution is “If project A is NOT done, then B must be done” xA+ xB ≥ 1
usually around the optimal linear solution
“Projects A or B, but not both, must be done” xA+ xB = 1
Step 3: For each integer point, plug into
the objective function to solve for their “Project A is done iff B and C are done” xB ≥ xA and xC ≥ xA

respective z-values. “If project A is done, then B can be done” xB ≥ xA

“If projects A and B are done, then C must NOT be done” xA + xB + xC ≤ 2


Note: LP-opt is always at least as good
as IP-opt (LP-opt is always equal or “Project A can be done only if B and C are done” xA≤ xB and xA≤ xC
better than IP-opt)
“Project A is done iff B and C are done” 2xA≤ xB+ xC and 1+xA ≥ xB+ xC

“If project A is done, then at least one of B and C must be xA ≤ xB + xC and xB + xC ≤ 2 - xA


done, but NOT both”

“If project A is NOT done, then both B and C are done” 2(1-xA) ≤ xB + xC

“If project A and B are done, then C must be done” 1+xC ≥ xA + xB

Downloaded by nidzo balling (nebokaca@lyft.live)

You might also like