Part Cutting Optimization With Bat Algorithm

Summary
The main theme of this study is to optimize the cuts in different sizes and shapes from a main material with Matlab software, inspired by the structure used during hunting.

I. INTRODUCTION

Part cutting optimization, also known as firing-less cutting optimization, is to be able to cut with more than one part in multiple dimensions on a plane that is considered two-dimensional in the optimization type. The purpose of the cuts is to keep them as close to each other as possible with the least amount of parts possible and the least distance between the parts to be cut.

II. WRITTEN ALGORITHM

The Bat Algorithm is an optimization algorithm inspired by the direction and distance detection behavior in which a cismin is obtained by using the echo of the sound called the echolocation of the yarns. It is proposed as an optimization algorithm based on herd intelligence by Yang in 2010. [1]

All living creatures (including dolphins, whales, gray rats and some species of birds) using echolocation, including wounds, emit a certain frekans signal. These signals are often beyond the limits of sound (about 20 kHz) that humans can perceive. Yarcs, using their echolocation system to determine their prey's location, communicate with each other, detect all kinds of objects even in completely dark environments, move without hitting them and distinguish any insects in motion. They can detect hunts and obstacles around them. In the context of the values, the bat changes position again to get closer to the target. A bat algorithm according to Yang is based on the following rules [2].

Rules

1- All wounds detect the location of the prey by ekolocation.

2- The velocity (v), position (x), frequency (f), wavelength (r) and sound output (A) of each wound are values.

3- They can adjust the wave length and sound output.

Location Update According to Ekolocation System

fi: The frequency value produced by taking into account the rastity of the minimum and maximum frequency range that a yarn can produce.

vi: The current subject and the generated frequency (fi) of a yarn is the new speed value

xi: The new position obtained by adding a new speed value to the former position of the knife.

II. OPTIONALIZING PARTS CUTTING

The following steps have been followed in practice in order to bring the cut parts to the most suitable positions and to the most suitable point on the base material.

Start-up phase of application

5 different sizes, different shapes (3 rectangles, 2 circles), and random pieces to be cut in different positions are created. Raw material area is defined as 30x30 piece.
The area covered by these parts on the main material was drawn with a virtual rectangle and the area account was made. The most appropriate value for the first step was selected.
The distance from the bottom left corner of this drawn virtual rectangle to the bottom left corner ((0,0) of the main material) is calculated. This distance was chosen as the most appropriate value for the first step.

Iteration phase

Five random population values ​​were determined for our bat algorithm.
The fitness value according to the objective function of the population values ​​is taken as the initial value for the first iteration.
Random positions were tried among the parts to be cut to try to find more suitable values.
Outcome Phase

The smallest value of the area covered by the parts cut during the interlace and the distance from the corner point ((0,0 coordinate)) on the base material is chosen as the appropriate value.

IV. PURPOSE FUNCTION

Our aim function is to optimize the parts in different dimensions which will be cut from the main part, to the least corner distance between them and to the corner point on the main part.

Optimization of the same parts to the corner point on the main part takes place via the bat algorithm, while the cut-off parts reach the appropriate values ​​according to the old optimize value depending on the rastality. So if we act on these rules, our goal is the problem of minimization.

Purpose Function of Cutting Parts

UAD: Appropriate Field Value

YAD: New Field Value

If (UAD> YAD)

UAD = YAD

Purpose Function on Main Part

KPA: The Area Covered by All the Fractures to Cut

OO: Most Point of Main Part ((0,0) Corner Coordinate)

KOS: Distance of KPA to OO (according to Euclidean method [3])

UU: Eligible Distance

If (UU> KOS)

U = KOS

V. RESULT

The high number of iterations resulted in a positive result in finding a more appropriate solution. But it has increased the transaction time.

VI. RESOURCES

[1] Xin-She Yang, A New Metaheuristic Bat-Inspired Algorithm 2010 University of Cambridge.

[2] A New Metaheuristic Bat-Inspired Algorithm![]

[3] Euclidean distance, https://en.wikipedia.org/wiki/Eklid_destination