Prediction of Capacity Utilization Rate for Turkey Using Adaptive Neuro-Fuzzy Inference System With Particle Swarm Optimization and Genetic Algorithm

Introduction

The manufacturing industry's place in the country's economy is one of the most fundamental determinants of a country's development process. The shares of the sectors in the economy also change according to the development status of the countries. While the sectors based on natural resources such as agriculture have an important share in the economy with the industrial revolution, as the countries' development rate increased, the agricultural sector was replaced by the manufacturing sector.

It is crucial to determine the production capacity and production planning for the continuity of an enterprise. Production capacity is the amount of production a business creates using the production factors rationally in a specific time. Long-term demand forecasts are needed to determine this production capacity. Conditions that create uncertainty regarding future demand and production costs obtained by long-term demand forecasts make capacity planning activities difficult.

Capacity utilization rate (CUR), on the other hand, gives the ratio of the production amount realized in the production unit to the highest amount that can be produced in a certain period. Due to the high share of the manufacturing industry in the industrial sector, this sector was utilized in measuring capacity utilization. The latest published capacity utilization rate report for Turkey belongs to May 2021. The report's numerical data show that the manufacturing industry's capacity utilization rate decreased by 0.6 points compared to the previous month and fell to 75.3 percent in May 2021. In Figure 1, seasonally adjusted capacity rate and capacity utilization rate values for Turkey are given for the years 2013-2021.

Figure 1.

CUR and Seasonally Adjusted CUR (%)

(Source: (CBTR, 2021))

According to Figure 1, the CUR decreased with a rapid acceleration between January and July of 2020. The declaration of the COVID-19 epidemic, which started in China in December 2019 as a pandemic by the World Health Organization (WHO) in March 2020, caused severe cuts in production, and the epidemic shook the manufacturing sector significantly. The effect of the epidemic on the manufacturing sector is seen in Figure 1.

Various methods are employed to find the CUR in the manufacturing sector in previous studies. These studies can be divided into two as statistical and artificial intelligence-based models. In addition, some studies conducted to investigate the manufacturing sector's capacity utilization rate with different methods such as (Ay, 2019; Dotsey & Stark, 2005; Eren Şenaras, 2016; Lima & Malgarini, 2016; Sen Rimo & Chai Tin, 2017; Sinan, 2020; Xu, 2019; Yu & Shen, 2020; Zia Ullah et al., 2017).

Previous studies on the capacity utilization rate have been summarized below. In these studies, prediction models were developed, and in others, indicators affecting CUR were evaluated.

Ay (2019) analyzed CUR and real sector confidence index data with the Granger Causality test. In this study, it is tested whether the use of capacity affects the price index. As a result of the research, a bidirectional causality relationship was found between the two variables. This shows that the variables have long-term interaction. (Ay, 2019).

Lima and Malgarini (2016) suggested estimating the output gap by examining the Brazilian manufacturing capacity utilization rate with the survey technique. System efficiency can be measured by observing the capacity utilization rate. If the output gap is determined well, an efficient decision support system will be created for the Brazilian business cycle. As a result, it is suggested that the survey technique can be used as a method to estimate the output gap with real-time GDP data set (Lima & Malgarini, 2016).