Question

Difficulty: HardOptimization and Bounded Constraints

An analytics firm processes data batches using two types of cloud computing instances: High-Memory Instances (MM) and High-Compute Instances (CC).

- Processing Capacity: Each High-Memory Instance processes 120120 thousand transactions per hour, and each High-Compute Instance processes 200200 thousand transactions per hour. The workload requires a total processing rate of at least 1,3201,320 thousand transactions per hour.
- Instance Availability: At most 1010 High-Memory Instances (M10M \le 10) and at most 66 High-Compute Instances (C6C \le 6) are available. At least 11 instance of each type must be used (M1M \ge 1 and C1C \ge 1).
- Load Balancing Constraint: To ensure infrastructure stability, the number of High-Memory Instances cannot exceed twice the number of High-Compute Instances (M2CM \le 2C).
- Operating Costs: High-Memory Instances cost $18\$18 per hour each, while High-Compute Instances cost $25\$25 per hour each.

Select the number of High-Memory Instances (MM) and the number of High-Compute Instances (CC) that minimize the total hourly operating cost while satisfying all operational requirements.

  • Number of High-Memory Instances (MM)1
  • Number of High-Compute Instances (CC)6

Answer

The minimum cost is achieved by selecting 1 High-Memory Instance (M=1M = 1) and 6 High-Compute Instances (C=6C = 6).
To minimize the total hourly cost subject to 3M+5C333M + 5C \ge 33, M2CM \le 2C, 1M101 \le M \le 10, and 1C61 \le C \le 6, testing all feasible integer boundary points shows that C=6C = 6 allows M=1M = 1 (since 3(1)+5(6)=333(1) + 5(6) = 33 and 12(6)1 \le 2(6) holds), producing the lowest cost of $168\$168. Thus, M=1M = 1 and C=6C = 6 are the optimal selections.

Step-by-Step Solution

1
Set up the objective function and mathematical constraints from the problem stem.
Objective: Minimize Cost=18M+25C\text{Cost} = 18M + 25C.
Constraints:
1) Throughput: 120M+200C1320    3M+5C33120M + 200C \ge 1320 \implies 3M + 5C \ge 33.
2) Availability: 1M101 \le M \le 10 and 1C61 \le C \le 6 (integers).
3) Load balancing: M2CM \le 2C.
Formulating the inequalities defines the boundary of feasible integer pairs (M,C)(M, C).
2
Test candidate integer boundary values for C{1,2,3,4,5,6}C \in \{1, 2, 3, 4, 5, 6\}.
- For C=1,2C = 1, 2: Maximum M2CM \le 2C yields 3(2C)+5C=11C3(2C) + 5C = 11C, giving max 1111 (for C=1C=1) and 2222 (for C=2C=2), both <33< 33 (infeasible).
- For C=3C = 3: M6M \le 6. 3M335(3)=18    M63M \ge 33 - 5(3) = 18 \implies M \ge 6. Min M=6M = 6. Cost =18(6)+25(3)=108+75=183= 18(6) + 25(3) = 108 + 75 = 183.
- For C=4C = 4: M8M \le 8. 3M335(4)=13    M53M \ge 33 - 5(4) = 13 \implies M \ge 5. Min M=5M = 5. Cost =18(5)+25(4)=90+100=190= 18(5) + 25(4) = 90 + 100 = 190.
- For C=5C = 5: M10M \le 10. 3M335(5)=8    M33M \ge 33 - 5(5) = 8 \implies M \ge 3. Min M=3M = 3. Cost =18(3)+25(5)=54+125=179= 18(3) + 25(5) = 54 + 125 = 179.
- For C=6C = 6: M10M \le 10. 3M335(6)=3    M13M \ge 33 - 5(6) = 3 \implies M \ge 1. Min M=1M = 1. Cost =18(1)+25(6)=18+150=168= 18(1) + 25(6) = 18 + 150 = 168.
Evaluating minimal feasible MM for each allowed CC identifies all vertex and boundary candidates.
3
Compare total cost across all valid candidate combinations.
The candidate pair (M=1,C=6)(M=1, C=6) yields the minimum operating cost of $168\$168, while satisfying 120(1)+200(6)=13201320120(1) + 200(6) = 1320 \ge 1320 and 12(6)1 \le 2(6).
Comparing all valid boundary points proves that M=1M=1 and C=6C=6 minimize cost under all constraints.

Key Concept

Optimization under Bounded Constraints
Rate this question