1

Given $S$ and a set of $n$ slices of salami ${C_1, C_2, ..., C_n}$, what is the most efficient way to place the slices of salami on $S$ such that the total area of $S$ covered by the slices of salami is maximized?

Efforts in solving the problem:

One possible approach to solving this problem is to use optimization techniques to maximize the total area covered by the slices of salami on the slice of sandwich. One possible optimization problem formulation is:

$$\max_{x_1, x_2, ..., x_n} \sum_{i=1}^{n} x_i A_i$$ subject to $$\sum_{i=1}^{n} x_i = 1$$ where $x_i$ is a binary variable indicating whether slice $C_i$ is placed on the sandwich, $A_i$ is the area of slice $C_i$, and the constraint ensures that exactly one slice of salami is placed on the sandwich.

This is a MILP problem and can be solved using various optimization solvers. However, this formulation assumes that the slices of salami are non-overlapping, which may not be the case in practice.

What would be a more realistic and practical optimization problem formulation for the placement of sliced salami on a slice of sandwich, taking into account the possibility of overlapping slices? How can we efficiently solve this problem?

rumathe
  • 438

1 Answers1

1

$$\max_{x_{ij}, p_i, q_i} \sum_{i=1}^{n} \sum_{j=1}^{m} x_{ij} A_{ij}$$ subject to $$\sum_{j=1}^{m} x_{ij} \leq 1 \quad \forall i$$ $$\sum_{i=1}^{n} \sum_{j=1}^{m} x_{ij} A_{ij} \leq A_S$$ $$A_{ij} \leq A_{C_i} \cdot x_{ij} \quad \forall i, j$$ $$0 \leq p_i \leq 1 \quad \forall i$$ $$0 \leq q_i \leq 1 \quad \forall i$$

where:

$x_{ij}$ is a binary variable indicating whether slice $C_i$ is placed on the sandwich with configuration $j$,
$p_i$ and $q_i$ represent the proportion of the maximum single straight cut allowed on slice $C_i$,
$A_{ij}$ is the area of slice $C_i$ in configuration $j$,
$A_S$ is the area of the slice of sandwich $S$,
$n$ is the number of slices of salami, and
$m$ is the number of possible configurations for each slice.

Can be solved efficiently using Branch-and-Bound method or the Outer Approximation algorithm.

rumathe
  • 438