**2.2 Metaheuristics**

*Textile Industry and Environment*

The consumption of water and energy was recorded from the beginning till the end of the study. Seven operations were evaluated as installing the flowmeters, semi-countercurrent rinsing, reusing the wastewater after treatment, recovering and reusing the wash water from mercerization, reusing the concentrate stream of reverse osmosis plant in sanitary, minimizing the water consumption during regeneration of ion exchangers used in water softening, and reusing the cooling waters in the production process. After implementation of the BAT actions, total water and energy consumption per kg fabric were reduced by 29.5 and 9%, respectively. As a result, saving from the expenses was roughly estimated as EUR 987,000 annually, while total amount of investment was EUR 45,000. There is no information given about the reduction of the amount of wastewater as a result of improvement. Yukseler and coworkers [27] tried to implement cleaner production to denim manufacturing textile mill in Turkey. The BAT methodology was followed to reduce the amount of consumed water and wastewater generation through the characterization of the wastewater and selection of the wastewater streams to reuse in the process. Selected BAT actions were reusing of wastewaters in the dyeing process after the treatment, recovering of caustic from alkaline finishing wastewaters, reusing of biologically treated composite mill effluent after membrane process, minimizing the wash water consumption in the water softening plant, reusing the concentrate stream from reverse osmosis plant, and reducing water consumption by countercurrent washing in dyeing and finishing processes. As a result, the reduction

in the total specific water consumption was evaluated as 30%.

of wastewater with toxicity analyses, and monitoring of flue gas emissions.

A LIFE funded project BATTLE (05 ENV/IT/000846) was proposed to evaluate the applicability of BAT such as those described in the textile reference documents (BREFs) for the implementation of the European Directive IPPC 96/91/CE to small-medium enterprises (SMEs) in terms of technical and economical feasibility of water recycling for European textile sector [30]. In the project, a prototype

Ozturk and coworkers studied on the improvement of a wool and acrylic fiber production mill in Turkey to a cleaner production by using BAT measures [28]. Suggested BAT actions were reuse of wastewaters from wool yarn softening, LP-VP printing machines and acrylic yarn washing, machinery modifications, reuse of steam condensate, and good management practices. Additionally, replacement of 12 toxic chemicals with biodegradable ones and installing an automatic dosage system were suggested in order to reduce the chemical load in the wastewaters. Energy saving precautions were determined as the implementation of energy recovery systems for high-temperature wastewater flows and flue gas streams, process monitoring control, and various machinery optimizations. After the mass balance calculations, it was estimated that all of the implementations could reduce total water consumption by 35–65%, total energy consumption by 70%, chemical load by 31%, and waste generation by 5–10%. The payback period of the installations was estimated as 4 years. Kalliala and Talvenmaa investigated the major six textile manufacturers in Finland considering environmental effect of wet processing and suggested appropriate actions of BAT [29]. The study was found especially important since all the industrial manufacturers discharge their wastewater to municipal sewage treatment plants under a strict control in Finland. Energy, water, and chemical consumption data of the processes were collected from the process statistics of the companies. Energy consumption was evaluated between 55 and 152 MJ/kg product, while water consumption was between 144 and 380 L/kg product, and the CO2 emission was found between 3484 and 8937 g/ kg product. A detailed chemical consumption table was also prepared for the study. As a result, suggested BAT actions were planned as the application of automatic dosing of chemicals and dyes, effective use of equipment capacity, recycling and monitoring of process water and energy used, recovery and purification of process liquor, monitoring

**4**

Using nonconventional optimization methods in the textile industry by considering both the delivery date and environmental issues is a quiet new area in the literature. Although the major part of these studies is accumulated on the scheduling of the dyeing process, novel optimization and decision-making algorithms (AI—Artificial Intelligence) have a huge potential in a large area of textile production such as cotton grading, yarn CPS (count strength product) prediction, yarn grading, fabric colorfastness grading, fabric comfort, and fabric inspection [31].

The studies on the nonconventional optimization methods can be classified into two groups: studies on the development of hybrid metaheuristic algorithms and the studies on the application of the genetic algorithm in real processes.

## *2.2.1 Developed hybrid algorithms*

Huynh and Chien [4] worked on the parallel batch processing machine scheduling with sequence-dependent setup, arbitrary job size, different due date, and incompatible job family. They proposed a multi-subpopulation genetic algorithm with heuristics embedded (MSGA-H) to improve batching and scheduling simultaneously. The validity of the algorithm was tested by an empirical study with data supplied from a textile dyeing manufacturing in Taiwan. The results have shown the practical viability of the proposed MSGA-H. The reduction of used water and wastewater were not reported in this study.

In another study on the textile dyeing process [20], parallel machine scheduling problem with environmental requirements and tardiness were solved by generating a multi-objective genetic algorithm with tabu-enhanced local search (MOGA-TIG). Three objective functions were defined to obtain a sustainable schedule: the number


#### **Table 1.**

*The reduction values in the water consumption, energy consumption, wastewater, and greenhouse gas generation.*

#### *Textile Industry and Environment*

of setups to be minimized to reduce the water consumption and use of detergent caused by the changing of the job, the utilization of the machines to be maximized to prevent the pollution caused by ineffective usage of the batch capacity, and tardiness that may cause penalty on late completion of the jobs to be minimized. When the performance of the algorithm was compared with the latest multi-objective evolutionary algorithms in the literature, it was concluded that the suggested algorithm gives a satisfactory solution in a short time. However, since the algorithm was not applied on a real industrial process, there was no data available for the reduction of the used water and dye in this study.

Zhang and coworkers studied on a similar process with the previous study, and they approached to the problem as bi-objective optimization model [21]. The objectives of the study were to reduce delivery tardiness cost and pollutant emission caused by the cleaning operation of setups before each job. Multi-objective particle swarm optimization algorithm enhanced by problem-specific local search

**7**

*Sustainable Production Methods in Textile Industry DOI: http://dx.doi.org/10.5772/intechopen.84316*

*2.2.2 Application of the genetic algorithms*

lessens the amount of wastewater.

**3. Conclusion**

approach.

the other enterprises.

(MO-PSO-L) technique was utilized to solve the problem. The algorithm was verified on computational experiments by using simplified realistic production data and also compared with the universal multi-objective optimizer NSGA-II and multiobjective imperialist competitive algorithm (MOICA). The performance of the algorithm was measured by using three indicators: the overall nondominated vector generation (ONVG), the C-metric (CM), and the Tan's spacing (TS). It was possible to conclude that MO-PSO-L algorithm found more Pareto solutions and became superior to other algorithms in terms of uniform distribution of the solutions.

Zhou and coworkers [22] used the genetic algorithm to schedule a dyeing process. Scheduling objectives are delivery time of the products, complete filling of the dyeing vessel, putting the same type, color, depth and production process of cloth in the same dyeing vessel, and sorting the depth of color such as from light to dark. Genetic algorithm was started with assigning all orders to different dye vats and producing a number of individuals. Each individual was checked to meet the delivery date. Then, fitness function was computed for the population. After the crossover and mutation steps, next generation was chosen according to the parent and child fitness. The procedure repeated itself until reaching the maximum number of evolutionary. The algorithm was verified on the data supplied from an enterprise in China. It was reported that the optimization of the schedule reduced the freshwater consumption about 18.4–21%, also resulting less wastewater. The flowchart of a typical genetic algorithm is shown in **Figure 1**.

Jiang and coworkers applied genetic algorithm to the data obtained from a real process in China to optimize production schedule [23]. The aim of the study was to reduce the production time and the volume of freshwater consumed in the dyeing operation. The results showed that the optimized schedule could reduce the production time as 10–15% while reducing the volumes of freshwater and wastewater as 20–30 and 20%, respectively. Additionally, it was found that adding alternative production lines (from 1 to 4) in the process shortens the total production time and

Textile industry, which is dominated by small and medium enterprises (SMEs), has a wide variety of products with different colors, fabrics, and fabric types by its nature. Additionally, wet processes consume large amount of water, energy, and chemicals, which are expensive to separate in the treatment processes. In the last decades, the studies with environmental and financial considerations in the textile industry have been increased, and they offered cleaner production approach, such as best available techniques (BAT) referred by the European Council and the nonconventional (metaheuristic) optimization methods, instead of end-of-pipe

From the literature, it was revealed that BAT actions offer a substantial water and energy savings up to 65 and 70%, respectively. However, many of the studies involve potential implementation results, not the real ones. Therefore, more implementation studies should be conducted on actual processes in order to encourage

Through the age of Industry 4.0, the enterprises, which use intelligent techniques and effective planning, will survive. Consequently, the studies on the optimization of the dyeing process scheduling have become more complex and

**Figure 1.** *Flowchart of the genetic algorithm [32].*

*Sustainable Production Methods in Textile Industry DOI: http://dx.doi.org/10.5772/intechopen.84316*

*Textile Industry and Environment*

the used water and dye in this study.

of setups to be minimized to reduce the water consumption and use of detergent caused by the changing of the job, the utilization of the machines to be maximized to prevent the pollution caused by ineffective usage of the batch capacity, and tardiness that may cause penalty on late completion of the jobs to be minimized. When the performance of the algorithm was compared with the latest multi-objective evolutionary algorithms in the literature, it was concluded that the suggested algorithm gives a satisfactory solution in a short time. However, since the algorithm was not applied on a real industrial process, there was no data available for the reduction of

Zhang and coworkers studied on a similar process with the previous study, and they approached to the problem as bi-objective optimization model [21]. The objectives of the study were to reduce delivery tardiness cost and pollutant emission caused by the cleaning operation of setups before each job. Multi-objective particle swarm optimization algorithm enhanced by problem-specific local search

**6**

**Figure 1.**

*Flowchart of the genetic algorithm [32].*

(MO-PSO-L) technique was utilized to solve the problem. The algorithm was verified on computational experiments by using simplified realistic production data and also compared with the universal multi-objective optimizer NSGA-II and multiobjective imperialist competitive algorithm (MOICA). The performance of the algorithm was measured by using three indicators: the overall nondominated vector generation (ONVG), the C-metric (CM), and the Tan's spacing (TS). It was possible to conclude that MO-PSO-L algorithm found more Pareto solutions and became superior to other algorithms in terms of uniform distribution of the solutions.
