Open Access is an initiative that aims to make scientific research freely available to all. To date our community has made over 100 million downloads. It’s based on principles of collaboration, unobstructed discovery, and, most importantly, scientific progression. As PhD students, we found it difficult to access the research we needed, so we decided to create a new Open Access publisher that levels the playing field for scientists across the world. How? By making research easy to access, and puts the academic needs of the researchers before the business interests of publishers.
We are a community of more than 103,000 authors and editors from 3,291 institutions spanning 160 countries, including Nobel Prize winners and some of the world’s most-cited researchers. Publishing on IntechOpen allows authors to earn citations and find new collaborators, meaning more people see your work not only from your own field of study, but from other related fields too.
To purchase hard copies of this book, please contact the representative in India:
CBS Publishers & Distributors Pvt. Ltd.
www.cbspd.com
|
customercare@cbspd.com
The Concept of Supply Chain Management (SCM) has been paid much more attention over the past decades. As one of the essential components of a supply chain, warehousing is valued because of the following major functions: smoothening the material flow; accomadating variability influenced by factors such as product seasonality or transportation scheduling; ensuring proper inventory level by product consolidation; guaranteeing the operation within high tolerances of speed, accuracy and lack of damage (Frazelle, 2002; Christopher, 2005; Harrison & van Hoek, 2005; Baker, 2007; Gu et al., 2007).
According to (Bernardy & Scherff, 1998), all the activities involved in a warehouse can be described by processes and are characterized by entailing a large number of differing, but interdependent sub-processes and many complex influential factors. Since there are diverse functional processes within which different combinations of influencing factors exist, the throughput capacity of the warehouse may be strongly affected, especially when the staffs at the operation level always keep different views upon process parameter settings based on their personal experiences. Hence it is essential to find out the optimal factor settings for the compound functional processes regarding the experts’ knowledge so as to make the right strategy, and finally obtain satisfying warehouse operation.
World has witnessed the soaring use of Artificial Intelligence (AI) for operations management (OM) with the purpose of decision support (Kobbacy et al., 2007). Hybrid architecture has become a new field of AI research, in light of the development of the next generation of intelligent systems. Current research in this field mainly concentrates on the marriage of Genetic Algorithms (GA) and Fuzzy Logic (Feng & Huang, 2005; Lau et al., 2009). Exploring the similarities of the essential structures of these two knowledge manipulation methods is where intelligent decision support systems can possibly play an important role. However, such hybrid systems have not shown great significance in the warehousing sector.
This chapter aims to develop a Fuzzy-GA capacity decision model (FGCDM) to enhance rack efficiency in a One-Warehouse, N-Supplier warehouse by taking into consideration the performance metrics and various driving factors of the related processes. The hybrid framework is proposed to enable decision makers to formulate nearly optimal sets of knowledge-based fuzzy rules so as to identify better solutions for fully utilizing the warehouse capacity.
The supply chain encompasses a complex set of activities which require a collection of metrics to adequately measure performance (Caplice & Sheffi, 1995; Tompkins & Smith, 1998). (Bowersox & Closs, 1996) states three objectives for developing and implementing performance measurement systems: to monitor historical system performance for reporting, to control ongoing performance so that abnormal processes may be prevented, and to direct the personnel’s activities. A conceptual framework for measuring the strategic, tactical and operational level performance in a supply chain is proposed in (Gunasekaran et al., 2001), in which performance measures on warehousing and inventory in a SCM was emphasized. An activity-based approach for mapping and analyzing the practically complex supply chain network is identified in (Chan & Qi, 2003), which can be regarded as a primary step on measuring the performance of processes. (Lohman et al., 2004) points out that by means of local key performance indicators (KPIs), The measurement scheme should be developing at a organization-wide scale. The interplay between organizational experiences and new performance measurement initiatives is highlighted (Wouters & Sportel, 2005). Furthermore, the research work in (Angerhofer & Angelides, 2006) shows how the key parameters and performance indicators are modelled through a case study which illustrates how the decision support environment could be used to improve the performance of a collaborative supply chain. (Niemi, 2009) optimizes the warehousing processes and assesses the related management attributes, realizing the objective of improving the warehousing practices and adopting more sophisticated warehousing techniques supported by knowledge sharing.
In addition, trade-off phenomenon on variable settings is a crucial aspect in the process-oriented supply chain. Leung and Spiring (Leung & Spiring, 2002) have introduced the concept of the Inverted Beta Loss Function (IBLF), which is a further deduction of the Taguchi Loss Function (Taguchi, 1986) in the industrial domain, helping to balance the possible loss resulting from trade-offs generated from different combinations of performance measures involved.
2.2. AI-based decision support system
Much work has been conducted in machine learning for classification, whereas the motivation is to attain a discovery of high-level prediction. Artificial intelligence (AI) has been widely used in knowledge discovery by considering both cognitive and psychological factors. Genetic Algorithm (GA), one of the significant AI search algorithms, is widely used to perform a global search in the problem space based on the mechanics of natural selection and natural genetics (Holland, 1992; Gen & Cheng, 2000; Freitas, 2001).
Inspiring from all above, a Fuzzy-GA Decision Capacity Model is proposed for decision-makers to better select the proper warehousing strategies in terms of the corresponding performance metrics. The capacity will be evaluated by the rack utilization of the designated warehouse.
The proposed decision-supoort approach consists of two major processes: knowledge representation and knowledge assimilation, which are shown in Fig.1.
Figure 1.
The proposed decision-support framework.
In the first stage, the expertise of factor setting, which is represented by IF-THEN rules, is encoded as a string with fuzzy rule sets and the associated fuzzy membership function. The historical process data are also included into the strings mentioned above, contributing to the formulation of an initial knowledge population. Then in knowledge assimilation, GA is used to generate an optimal or nearly optimal fuzzy set and membership functions for the entitled performance indicators. Accordingly, it is necessary to set relative weights for them to aggregate the measurement results since there naturally contains essential fuzziness and ambiguity in human judgments.
Fig. 1 depicts the overview of the entire proposed knowledge-based framework, while the initial rules extracted from process knowledge base are used to form the initial population of the GA. Fig. 2 illustrates the data flow of the proposed capacity-optimizing model, indicating how the iterations envelop fuzzy rule mining, improving the quality of generated rule sets and streamlining the various functional processes in a single warehouse.
Figure 2.
Information flow of the proposed algorithm (Reference: Ho et al., 2008).
3.1. Problem formulation
Fuzzy-encorporated GA is proposed for capturing domain knowledge from an enormous amount of data. The proposed approach is to represent the knowledge with a fuzzy rule set and encode those rules together with the associated membership into a chromosome. A population of chromosomes comes from the past historical data and an individual chromosome represents the fuzzy rule and the related problem. A binary tournament, using roulette wheel selection, is used for picking out the best chromosome when a pair of chromosomes is drawn. The fitness value of each individual is calculated using the fitness function by considering the accuracy and the trade-off of the resulting performance measure setting, where the fitter one will remain in the population pool for further mating. After crossover and mutation, the offspring will be evaluated by the fitness function and the optimized solution will then be obtained.
The practitioners could freely select the specifically influential performance measures from a large pool of the candidate performance metrics based on the unique condition of the warehouse, leading to the optimized warehousing rack efficiency amongst all by comparing the weights.
3.2. Nomenclature
Table 1.
Nomenclature of the proposed algorithm.
Table. 1 above indicates the notations of the mathematical expressions involved in the proposed decision-support algorithm.
3.3. Chromosome encoding
Fuzzy concept is used to map the above linguistic decision rules into genes for GA optimization.
Definition 1: Ch={1,2,...,M}represents the index set of chromosomes where M is the total number of chromosomes in the population.
Definition 2: Gm×trepresents a gene matrix generated for the population where
Note that the decoding method of an element in the first sub-matrix (piv)m×b or second sub-matrix (dix)m×s of Gm×w to a linguistic variable is given by:
(i) 0: ignore, (ii) 1: low, (iii) 2: medium, and (iv) 3: high. For any row of the third sub-matrix (kiy)m×e ofGm×w, a group of six consecutive values ki(6ρ−5),ki(6ρ−4),ki(6ρ−3),ki(6ρ−2),ki(6ρ−1),ki(6ρ) in the matrix forms a single set Fpiv~={cpiv−wpiv,wpiv,cpiv,wpiv,cpiv+wpiv,wpiv} for process parameter pv whereρ=1,2,3,....... Also, for any row of the fourth sub-matrix (qiz)m×n ofGm×w, a group of six consecutive values qi(6ρ−5),qi(6ρ−4),qi(6ρ−3),qi(6ρ−2),qi(6ρ−1),qi(6ρ) in the matrix forms a single set Fdix~={cdix−wdix,wdix,cdix,wdix,cdix+wdix,wdix} for defect rate dx whereρ=1,2,3,....... For both two cases, there are totally 6 genes in the sets of membership functions shown in Fig. 3.
Figure 3.
Fuzzy membership functions of the influencing factors.
Fpiv~ consists of aggregated membership functions which relate to a fuzzy rule set is assumed to be isosceles-triangle functions.
cpivis the center abscissa ofFpiv~; wpivrepresents half the spread ofFpiv~.
In “cpiv”, “piv” indicates that the v-th feature test is included, while i specifies the order of all the condition levels of each feature test. For instance, cpi1stands for the center abscissa of the 1st process test, within the whole membership function matrix.
Definition 3: Bm×1denotes a random number matrix generated for selection and crossover where
Bm×1=(bi)m×1E4
bi=random[0,1]∀i∈Chm=ME5
.
Definition 4: Ch_c={1,2,.....,S}denotes the index set of the chosen chromosomes in the crossover where S is the total number of chosen chromosomes
Definition 5: G′m×windicates the gene matrix in which the Q chromosomes chosen in crossover are stored where
G'm×w=((p'iu)m×a(d'ix)m×b(k'iy)m×c(q'iv)m×z)E6
3.4. Fitness evaluation
To have a good set of process parameters, the genetic algorithm selects the best chromosome for mating according to the fitness function suggested below.
Fitness Funtion=accuracy with error rateE7
Accuracy=objects correctly matched within error range total number of objectsE8
Error rate (ε)=∑j=1mwj(yj−yj')22nE9
Each chromosome is evaluated by calculating its mean-square error for the error measurement. As each chromosome is represented as the fuzzy rule, the quality of the chromosome is then validated by comparing its defuzzified output with the actual output of the test samples. The centre of gravity (COG) is used as the defuzzification method to obtain the crisp values of the finished quality level.
3.5. Chromosome crossover
Crossover is a genetic operation aiming at producing new and better offspring from the selected parents, while the selection is determined by a crossover rate. The current crossover methods include single-point crossover, two-point crossover, multi-point crossover, uniform crossover, random crossover, etc. Uniform crossover is selected in this research.
3.6. Chromosome mutation
Mutation is intended to prevent all solutions in the population from falling into the local minima. It does this by preventing the population of chromosomes from becoming too similar to each other, which might slow down or even stop evolution. Mutation operation randomly changes the offspring resulting from crossover, given that the value of the mutation rate must range within 0 and 1. In our paper a bit-flip mutation is used.
3.7. Chromosome repairing
After the mutation and crossover in the two regions, some violations in the chromosome may occur. If the membership function is not in ascending order, the new offspring should be modified by exchanging the gene order in accordance with the definition ofFpiv~={cpiv−wpiv,wpiv,cpiv,wpiv,cpiv+wpiv,wpiv}.
The repairing is divided into two categories which are: the forward and backward repairing as illustrated in Fig.4(a) and Fig. 4(b).
Figure 4.
a) Sample chromosome of forward repairing. (b) Sample chromosome of backward repairing.
3.8. Chromosome decoding
Once the termination criterion is fulfilled, the decoding process will be implemented on the whole set of optimum chromosomes (Fig. 5). The optimum chromosomes decode into a series of linguistic fuzzy rule sets as shown in Table 2 and their associated membership functions which are stored in the repository for further investigation.
Condition part <IF"/ ( Warehousing Influencing Factors )
Consequent part <THEN"/ ( Rack Utilization )
Rule 1 : Process1 . Inventory cost is adjusted to medium AND Process2 . Backorder cost is adjusted to medium AND Process 3 . Maintenance cost is adjusted to high ... ... AND Process N .
Rack utilization of Drive-in is extremely low AND Rack utilization of APR is high AND Rack utilization of Double-deep is medium AND ...
Rule 2: Process1 .Inventory cost is adjusted to low AND Process4. Backorder cost is adjusted to high AND ... ... AND ProcessN +1.
Rack utilization of Drive-in is medium AND Rack utilization of Double-deep is extremely high
Table 2.
Table 2. Sample of generalized fuzzy rules obtained in the FGCDM.
Figure 5.
Sample of generalized fuzzy rules obtained in the FGCDM.
3.9. De-fuzzification
Once the termination criterion is fulfilled, the decoding process will be implemented on the whole set of optimum chromosomes. The optimum chromosomes decode into a series of fuzzy rule sets and their associated membership functions which are stored in the repository for further investigation.
The warehousing background for the simulation is of medium volumes (300 pallets/day throughput) and with 90 SKUs to be placed into the storage. The existing rack system include Block-stack, Drive-in, APR, Double deep and VNA. The evaluation criterion of the warehouse performance is mainly based on the utilization of the above racks.
In order to verify the proposed Fuzzy-GA capacity decision model (FGCDM), simulations on searching ability were carried out. Two different stochastic-based search methods, Simulated Annealing (SA) and Tabu Search (TS), were used for comparison with the proposed FGCDM approach. In this experiment, the historical data for supporting the warehousing operation and 30 performance indicators were used for the simulation. The results reported are all averaged over 10 independent runs. In each data set, the best (minimum) fitness value among the 10 simulation runs was documented for the comparison of each search technique mentioned above.
Number of runs
SA
TS
FGC DM
1
0. 822
0. 89
0.913
2
0.8 7
0. 9 2 3
0.892
3
0.9 1
0. 887
0.93
4
0. 762
0.7 81
0.795
5
0. 863
0.8 71
0.88
6
0. 836
0. 82
0.933
7
0. 816
0. 848
0.853
8
0. 902
0.8 33
0.892
9
0. 827
0. 911
0.958
10
0.8 42
0.8 92
0.884
Average
0.845
0.866
0 . 8 9 3
Table 3.
Best (Minimum) fitness values obtained by FGCDM, SA & TS.
Rack Utilizations (%)
Warehouse Rack Type
Model Result
Observed
Block-stack
91.8%
88.2%
Drive-in
91.2%
75.7%
APR
95.5%
96.1%
Double deep
89.7%
77.3%
VNA
93.1%
92.8%
Table 4.
Rack Utilization of Observed and Model Results.
Table 3 presents that ten independent runs of fitness values acquired by various search techniques using 30 performance indicators. According to the experiment, SA was the worst performer in all 10 independent runs and the proposed FGA approach achieved the smallest average object value at 0.893 in the maximization of rack utilization over the interval 0 to 1. Compared with the observed test data which are half-extracted from the historical records, our approach shows an overall better result in Table 4.
In this research, the design and implementation of a GA based process knowledge model, which embraces the fuzzy theory and genetic algorithm to achieve warehouse capacity improvement, has been introduced. Implementing the proposed methodology in the aspect of warehouse management through simulation has been successful. By incorporating the error measurement and complexity of process change into the fitness evaluation, the generalized fuzzy rule sets can be of less complexity and higher accuracy. An extension of different measures can also be included in order to improve the generalized rules. In the matter of generation of new fuzzy rules, the membership functions are assumed to be static and known. The proposed intelligent model can help the decision makers in the development and selection of the best warehouse design for the given application.
Other fuzzy learning methods should be considered to dynamically adjust the membership functions of various parameters to enhance the model accuracy. Future contribution of this endeavour goes to validation of the decision model to be launched in case companies.
The authors wish to thank the Research Committee of The Hong Kong Polytechnic University for the support of this research.
References
1.Al-KuzeeJ.MatsuuraT.GoodyearA.NolleL.HopgoodA. A.PictonP. D.BraithwaiteN. St. J.2004Optimization of plasma etch processes using evolutionary search methods with in situ diagnostics, Plasma Sources Science Technology, 134612622 .
2.AngerhoferB. J.AngelidesM. C.2006 IA model and a performance measurement system for collaborative supply chains, Decision Support Systems, 421283301 .
3.BakerP.2007An exploratory framework of the role of inventory and warehousing in international supply chains, International Journal of Logistics Management, 1816480 .
4.BernardyG.ScherffB.1998 SPOC- Process Modelling Provides On-line Quality Control and Predictive Process Control in Particle and Fibreboard Production. Proceedings of the 24th Annual Conference of IEEE Industrial Electronics Society, IECON’98, 31.08.-04.09., Aachen.
5.BowersoxD. J.ClossD. J.1996Logistical Management: the Integrated Supply Chain Process, Macmillan, New York, NY.
6.CapliceC.SheffiY.1995A review and evaluation of logistics performance measurement systems, The International Journal of Logistics Management, 616174 .
7.ChanF. T. S.QiH. J.2003Feasibility of performance measurement system for supply chain: a process-based approach and measures, Integrated Manufacturing Systems, 143179190 .
8.ChenM. Y.LinkensD. A.2004Rule-base self-generation and simplification for data-driven fuzzy models, Fuzzy Sets and Systems, 1422243265 .
9.ChiangT. C.HuangA. C.FuL. C.2007 Modeling, scheduling, and performance evaluation for wafer fabrication: a queueing colored Petri-net and GA-based approach, IEEE Transactions on Automation Science and Engineering, 33912918 .
10.ChristopherM.2005Logistics and Supply Chain Management, third ed. Pearson, Harlow.
11.CordonO.Del JesusM. J.HerreraF.1998Genetic learning of fuzzy rule-based classification systems cooperating with fuzzy reasoning methods, International Journal of Intelligent Systems, 1310-11 , 10251053 .
12.De la IglesiaB.PhilpottM. S.BagnallA. J.Rayward-SmithV. J.2003 Data Mining Rules Using Multi-Objective Evolutionary Algorithms, In: Proceedings of IEEE Congress on Evolutionary Computations, 315521559 .
13.FengX.HuangH.2005A fuzzy-set-based Reconstructed Phase Space method for Identification of Temporal Patterns in Complex Time Series, IEEE Transactions on Knowledge and Data Engineering, 175601613 .
14.FrazelleE.2002World-class Warehousing and Material Handling. McGraw-Hill, New York.
15.FreitasA.2001A survey of evolutionary algorithms for data mining and knowledge discovery. In: Advances in Evolutionary Computation. Springer- Verlag.
16.GenM.ChengR.2000Genetic algorithms and engineering optimization. New York: Wiley.
17.GomezJ.GonzalezF.DasguptaD.2002 Complete Expression Trees for Evolving Fuzzy Classifier Systems with Genetic Algorithms, In: Proceedings of the Evolutionary Computation Conference GECCO’02, 2002.
18.GopalanJ.AlhajjR.BarkerJ.2006 Discovering Accurate and Interesting Classification Rules Using Genetic Algorithm, In: Proceedings of the 2006 International Conference on Data Mining, 389395 . June 26-29, 2006.
19.GuJ.GoetschalckxM.McGinnisL. F.2007Research on warehouse operation: A comprehensive review. European Journal of Operational Research, 1771121 .
20.GunasekaranA.PatelC.TirtirogluE.2001Performance measures and metrics in a supply chain environment, International Journal of Operations & Production Management, 211 2, 7187 .
21.HarrisonA.van HoekR.2005Logistics Management and Strategy. second ed. Pearson, Harlow.
22.HasanzadeM.BagheriS.LucasC.2004 Discovering Fuzzy Classifiers by Genetic Algorithms, In: Proceedings of 4th international ICSC Symposium on Engineering of Intelligent Systems (EIS2004), 2004, Island of Madeira, Portugal.
23.HigginsonJ. K.BookbinderJ. H.2005 Distribution centres in supply chain operations. In: Langevin, A.L. & Riopel, D. (2005), Logistics Systems: Design and Optimization. Springer, New York, 6791 .
24.HoG. T. S.LauH. C. W.ChungS. H.FungR. Y. K.ChanT. M.LeeC. K. M.2008Development of an intelligent quality management system using fuzzy association rules, Industrial Management & Data Systems, 1087947972 .
25.HollandJ. H.1992Adaptation in Natural and Artificial Systems. Cambridge, MA: MIT Press.
26.IshibuchiH.YamamotoT.2002 Fuzzy rule selection by data mining criteria and genetic algorithms, In: Proceedings of Genetic and Evolutionary Computation Conference (GECCO 2002), 399406 , New York, July 9-13.
27.IshibuchiH.YamamotoT.2004Fuzzy Rule Selection by Multi-Objective Genetic Local Search Algorithms and Rule Evaluation Measures in Data Mining, Fuzzy Sets and Systems, 14115988 .
28.KobbacyK.VaderaS.RasmyM. H.2007AI and OR in management of operations: history and trends, Journal of the Operational Research Society, 581028 .
29.LauH. C. W.TangC. X. H.LeungB. P. K.LeeC. K. M.HoG. T. S.2009A Performance Tradeoff Function for Evaluating Suggested Parameters in the Reactive Ion Etching Process, IEEE Transactions on Systems, Man, and Cybernetics- Part A: Systems and Humans, 375758769 .
30.LeungR. W. K.LauH. C. W.KwongC. K.2003An expert system to support the optimization of ion plating process: an OLAP-based fuzzy-cum-GA approach, Expert Systems with Applications, 253313330 .
31.LeungB. P. K.SpiringF. A.2002The inverted beta loss function: properties and applications, IIE Transactions, 341211011109 .
32.LiT. S.SuC. T.ChiangT. L.2003Applying robust multi-response quality engineering for parameter selection using a novel neural-genetic algorithm, Computers in Industry, 501113122 .
33.LohmanC.FortuinL.WoutersM.2004Designing a performance measurement system: A case study, European Journal of Operational Research, 1562267286 .
34.LopesC.PachecoM.VellascoM.PassosE.1999Rule-Evolver: An Evolutionary Approach For Data Mining, In: Proceedings of the 7th International Workshop on Rough Sets, Fuzzy Sets, Data Mining, and Granular-Soft Computing, 458462 .
35.MendesR. R. F.VoznikaF.de B FreitasA. A.NievolaJ. C.2001Discovering Fuzzy Classification Rules with Genetic Programming and Co-Evolution, In: Principles of Data Mining and Knowledge Discovery (Proceedings of the 5th European Conference PKDD 2001)-Lecture Notes in Artificial Intelligence, Springer-Verlag.
36.MilfelnerM.KopacJ.CusF.ZuperlU.2005Genetic equation for the cutting force in ball-end milling, Journal of Materials Processing Technology, 164 165, 15541560 .
37.NiemiP.HuiskonenJ.KarkkainenH.2009Understanding the knowledge accumulation process-Implications for the adoption of inventory management techniques, International Journal of Production Economics, 1181160167 .
38.RadhakrishnanP.PrasadV. M.GopalanM. R.2009 Inventory Optimization in Supply Chain Management using Genetic Algorithm, International Journal of Computer Science and Network Security, 913340 .
39.SantosC. A.SpimJ. A.IerardiM. C. F.GarciaA.2002The use of artificial intelligence technique for the optimisation of process parameters used in the continuous casting of steel, Applied Mathematical Modelling, 261110771092 .
40.TangC. X. H.LauH. C. W.2008 A Fuzzy-GA Decision Support System for Enhancing Postponement Strategies in Supply Chain Management, In: Lecture Notes in Computer Science, 5361 Springer-Verlag Berlin Heidelberg, 141150 .
41.TaguchiG.1986Introduction to Quality engineering: Designing Quality into Products and processes. NY: Kraus, White Plains.
42.TengM.XiongF.WangR.WuZ.2004 Using genetic algorithm for weighted fuzzy rule-based system, In: Proceedings of Fifth World Congress on Intelligent Control and Automation, 2004, Hangzhou, China.
44.WoutersM.SportelM.2005The role of existing measures in developing and implementing performance measurement systems, International Journal of Operations & Production Management, 251110621082 .
45.ZhuF.GuanS. U.2004Ordered Incremental Training with Genetic Algorithms, International Journal of Intelligent Systems, 191212391256 .