E

Sp

I

anFcTppoprdetfddbdfe

s

1234567

atoe

O

S

e

1d

Applied Soft Computing 10 (2010) 959–962

Contents lists available at ScienceDirect

Applied Soft Computing

journa l homepage: www.e lsev ier .com/ locate /asoc

ditorial

pecial issue on optimisation methods & applications in decision-making

rocesses

ntroduction

Decision-making and optimisation are oft studied subjectsnd the scopes of these areas are vast. Decision-making is aatural phenomenon, and in nature is a self-adaptive process.urther, decision-making and optimisation techniques requireognisance of the issues of human values, ethics and activities.hese “human interferences” compound the inherent com-lexities of goal-oriented decision-making processes. One suchroblem is encountered while developing a robust large-scaleptimisation/decision-making unit that optimises the controllingarameters, under uncertainty and complex multi-objective crite-ia. This Special Issue brings together researchers working on theevelopment of decision-making and optimisation techniques fornhancing the performance of modern operational research sys-ems in the field of engineering, business and management. Theocus of the Special Issue is to present the current trends in theevelopment of optimisation techniques and their application inecision-making processes. A balanced set of new research haseen selected that addresses real world complexities in the arena ofecision-making and optimisation using MCDM/MADM methods,uzzy methods, hybrid methods, multi-objective programming,volutionary algorithms, ACO, PSO and DEA.

This special issue includes thirty contributions, grouped intoeven categories:

. Multi-criteria decision-making processes,

. Fuzzy decision-making and optimisation methods,

. Multi-objective programming and related issues,

. Genetic algorithms,

. Hybrid genetic algorithms,

. Swarm intelligence, and

. Meta-modelling & other heuristics.

Contributions mainly present either a novel approach applied toproblem or a state-of-the-art method utilised for solving indus-

rial management, logistics and supply-chain problems. A briefutline of the contributions included in this Special Issue is delin-ated below.

rganisation of the special issue

ection 1 – multi-criteria decision-making processes

The first paper contains a broad spectrum of human values,thics and activities in considering the structures of decisions in

568-4946/$ – see front matter © 2010 Published by Elsevier B.V.oi:10.1016/j.asoc.2010.07.006

order to serve the needs of decision makers. Saaty and Begicevicidentify three lists of human values and activities for decision-making purposes. The manuscript opens up several new researchagendas relating to aspects of ethics, values and human activitiesin decision-making.

The second paper raises issues of inconsistency in pair-wisereciprocal matrices of Analytic hierarchy Process (AHP). Theseissues lead Yuen to propose an objective hierarchy model – AnalyticHierarchy Prioritization Process (AHPP) – in order to approximatethe real priority vectors by selecting the most appropriate prioriti-zation operator from candidate alternatives. It has been argued thatAHPP is one of the methods to address the prioritization problem soas to make better decisions. It has been also indicated that the mostappropriate prioritization operator is dependent of the content ofthe reciprocal matrix of the AHP model.

In the third paper, Dey reports an integrated risk manage-ment framework, for managing project risks, in order to analyserisk across project, work package and activity levels. The paperintegrates AHP and risk map methodologies. The proposed frame-work suggests mitigation measures for technical, organisationaland environmental risk factors thereby providing the dynamic deci-sions during project planning phase. The paper also develops somenew research directions in the field.

Zammori in the fourth paper reports that, in addition to complexdecision-making procedures, both AHP and Analytic Network Pro-cess (ANP) can be used as forecasting models effectively. A globalview of AHP & ANP and its credibility in dealing with complex reallife decisions subjected to a diversity of influences is also illustrated.The paper examines the potential of the AHP and ANP models tohelp discern states and situations as well as to predict outcomes.One example uses AHP to predict the Democratic nominee in the2008 United States presidential election and the overall electionwinner. In another example, ANP predicts the market share for skiequipment.

The fifth paper of Bhattacharya et al. report a hierarchicalconcurrent engineering approach integrating AHP with QualityFunction Deployment (QFD) in combination with Cost Factor Mea-sure (CFM) for ranking and selecting candidate-suppliers undermultiple, conflicting-in- nature criteria environment. Engineeringrequirements and customer requirements governing the selectiondecision have been identified. The hierarchical QFD methodology

allows the Decision-Maker (DM) to rank the candidate-suppliersconsidering both CFM and the subjective factors. Experimentalvalidation of the methodology is conducted with design of experi-ments.

9 ompu

S

oPpfD

MthcunaHa

cao(tDaoreria

amodbms

fctrtdcw

stdtsmrtath

S

a

60 Editorial / Applied Soft C

ection 2 – fuzzy decision-making and optimisation methods

In the sixth paper, Sadi-Nezhad et al. assess the performancef traffic police centres using a fuzzy “Technique for Order ofreference by Similarity to Ideal Solution” (TOPSIS) method. Theroposed approach is based on the preference ratio and an efficientuzzy distance measurement for a Fuzzy Multiple Criteria Groupecision-Making Problem (FMCGDMP).

The seventh paper by Azadeh et al. proposes a revised Nicholas’ining Method Selection (MMS) technique. Inspired by Nicholas’

echnique they present a two-step fuzzy MCDM algorithm havingierarchical technical–operational model (HTOM) and hierarchi-al economical model (HEM). HTOM ranks the mining alternativessing fuzzy-AHP and HEM selects the most profitable of those alter-atives. The judgmental weights have been assigned to unsteadynd uncertain characteristics of mineral resources using Analyticierarchy Process (AHP). Trapezoidal fuzzy numbers are used toddress the uncertainty issues.

The eighth paper, by Hatami-Marbini et al., integrates the con-ept of the TOPSIS into a four-phase fuzzy data envelopmentnalysis (DEA) framework, based on the theory of displaced ideal, inrder to measure the efficiencies of a set of decision-making unitsDMUs) and rank them with fuzzy input–output levels. Two hypo-hetical decision-making units (DMUs), called the ideal and nadirMUs, are constructed and used as reference points to evaluateset of information technology (IT) investment strategies based

n their Euclidean distances from these reference points. The bestelative efficiency of the fuzzy-ideal DMU and the worst relativefficiency of the fuzzy-nadir DMU are determined and combined toank the DMUs. It has been claimed that the proposed frameworks generic, structured and comprehensive and can be applied tonalyse various DMU evaluation problems in fuzzy environments.

Chou et al., in the ninth paper, report an empty containerllocation problem by designing a two-stage mixed fuzzy decision-aking and optimisation programming model. The optimal volume

f empty containers at a port (fuzzy backorder quantity inventoryecision-making model) and re-distribution of empty containersetween ports (mathematical programming network model) toeet exporters’ demand over time using this approach is demon-

trated.In a second paper, Chou et al. formulate a two-stage combined

uzzy-MCDM and optimisation programming model for solving theontainer transportation demand split problem. The first stage ofhe model deals with the container transportation demand splitate by using the fuzzy-MCDM method. In stage two, an optimiza-ion mathematical programming network model is proposed foretermining the inland Origin-Destination (O-D) of import/exportontainers. The model is validated with a case study from Tai-anese ports.

The eleventh paper by Liu et al. reports that PROFITS – a softwareystem – can be used for serving modern international containerransportation. The paper considers three modules of PROFITS: theemand forecasting, stowage planning and shipping line optimisa-ion. The system constructs problem models and uses exponentialmoothing, regression analysis, neural network, linear program-ing, genetic algorithm and sequence alignment methods to solve

elevant issues. The system relates to the whole links of maritimeransportation and the software uses various features of DSS by cre-ting a framework for forecasting and optimisation so as to providehe decision-support. It has been claimed that the proposed systemelps companies significantly to increase their profits.

ection 3 – multi-objective programming and related issues

In the twelfth paper, Ghoseiri and Ghannadpour present a modelnd solution for a multi-objective Vehicle Routing Problem with

ting 10 (2010) 959–962

Time Windows (VRPTW) using goal programming and genetic algo-rithms. VRPTW involves the routing of a set of vehicles with limitedcapacity from a central depot to a set of geographically dispersedcustomers with known demands and predefined time windows.The algorithm is applied to solve the benchmark Solomon’s 56VRPTW 100-customer instances.

In the next paper, Kang and Lee formulate a colour filterinventory replenishment problem as a fuzzy multiple-objectiveprogramming for a Thin Film Transistor-Liquid Crystal Display(TFT-LCD) manufacturing company considering storage space, yieldrate, quantity discounts and multiple suppliers. The next phaseof their model utilises Fuzzy Analytic Hierarchy Process (FAHP)and Extent Analysis Method (EAM) in order to incorporate experts’opinions into the decision-making process.

Moura et al. report the use of a “Multi-Objective GeneticAlgorithm (MOGA) using Pareto-ranking” as a means of compar-ing solutions across the multiple objectives in order to solve aroute planning task for multiple Autonomous Underwater Vehi-cles (AUV). The main objectives of the paper are the minimisationof the total travel distance and the maximisation of the number ofsamples so as to prevent the collisions between AUVs.

Section 4 – genetic algorithms

Issues relating to the field of Genetic Algorithm (GA) begin withthe contribution of Gulek and Toroslu. They report a Maximum-Weighted Tree Matching Problem (MWTMP) for the assignmentof tasks to the divisions of a hierarchical organisation and proposea GA solution. In the GA solution, Kuhn-Munkres algorithm is usedfor the optimal assignment.

Ahrari and Atai, in the sixteenth paper, introduce an evolution-ary algorithm – Grenade Explosion Method (GEM) – for optimisingreal-valued bounded black-box optimisation problems. GEM is ableto find the global minimum while surrounded by many local min-ima, all global minima of functions with multiple global minimaand the high-fitted local minima with a probability which stronglydepends on the fitness of that local minimum. The reported algo-rithm finds the global minimum of multimodal functions withoutbeing trapped in local minima. It has been illustrated that thealgorithm converges to the global minima faster than other evo-lutionary methods, e.g., GA and Artificial Bee Colony (ABC).

The seventeenth paper, by El-Mahdy et al., introduce the appli-cation of GA in determining the optimum size of pipes for thepredefined topology-based looped natural gas pipe network usedin natural gas applications. The goal is to minimise the networkcost. Network parameters are optimised utilising GA. The discretenature of decision variables, (viz. pipe diameters), and hard and softconstraints are considered in the algorithm.

Section 5 – hybrid genetic algorithms

A hybrid algorithm, called Self-Learning Genetic Algorithm(SLGA), is presented by Hakimi-Asiabar et al. This application ofSLGA is aimed at deriving optimal operating policies for the Karoon-Dez multi-reservoir system, a surface water resource in Iran, wherethere are three objectives to be addressed. Their SLGA uses aSelf-Organizing Map (SOM) algorithm, a neural network which iscapable of learning, and a Variable Neighborhood Search (VNS)algorithm – a local search efficiency enhancer – in order to adda memory to the genetic algorithm and improve its local searchaccuracy.

Application of a hybrid genetic algorithm in the field of dynamicand automated classification of patent documents and a patentsearching system is addressed in the eighteenth paper of thisSpecial Issue, where Wu et al. present an AI-aided patent decision-making process. The methodology integrates an expert screening

ompu

aMedted

atmnmmuis

epedseatro

B2Daifov

rtpbfstmtuood

S

aafraacttm

Editorial / Applied Soft C

pproach with a Hybrid Genetic Algorithm-based Support Vectorachine (HGA-SVM) model. This approach optimises all the param-

ters of SVM in developing the patent classification system withesired accuracy and generalisation ability. The effectiveness ofhis approach is demonstrated in an experiment conducted on anxpert’s patent document searching history having 234 semicon-uctor equipment components patent documents.

Strnad and Guid, in the twentieth paper, present a fuzzy-geneticnalytical model to facilitate an optimal formation of a projecteam. Based on existing quantitative approaches, the proposed

odel adds modelling enhancements, viz., derivation of person-el attributes from dynamic quantitative data, complex attributeodelling and handling of necessary over-competency. The opti-isation part of the selection of multiple project teams’ process

nder conflicting-in-nature criteria involves a special adaptation ofsland GA with mixed crossover where the fitness of the commonolution is used to drive the selection within the islands.

In the twenty-first paper, Liao proposes two hybrid differ-ntial evolution algorithms for engineering design optimisationroblems. The first hybrid algorithm enhances a basic differentialvolution algorithm with a local search operator; random walk withirection exploitation, to strengthen the exploitation ability. Theecond adds a second meta-heuristic, harmony search, to coop-rate with the differential evolution algorithm so as to producedesirable synergetic effect. It has been established that both of

he hybrid algorithms outperform the differential evolution algo-ithms. The cooperative hybrid algorithm overall outperforms thether hybrid algorithm having local search.

The twenty-second paper by Chiang et al. proposes a 2-Optased Differential Evolution (2-Opt DE) algorithm inspired by the-Opt Algorithms in order to accelerate global optimization usingE. The mutation schemes of 2-Opt DE, DE/2-Opt/1 and DE/2-Opt/2re substituted for mutation schemes of the original DE. A compar-son of 2-Opt DE and DE is provided using a set of 19 benchmarkunctions for experimental verification. This shows that 2-Opt DEutperforms the original DE in terms of solution accuracy and con-ergence speed.

The contribution of Komal et al. develops a Genetic Algo-ithms based Lambda–Tau (GABLT) approach, a hybrid technique,o estimate the Reliability, Availability and Maintainability (RAM)arameters of complex and repairable industrial systems. Thisegins with analysing the systems’ behaviour on the basis of pastailure and repair data and improving the performance by adoptinguitable maintenance strategies. Expressions for the RAM parame-ers of the system are obtained by using the traditional Lambda–Tau

ethodology. GA is used to compute the parameters in the form ofhe triangular fuzzy membership functions. A general RAM-Index issed for post-RAM analysis so as to rank the systems’ componentsn the basis of their performance. It has been claimed that GABLTptimises the uncertainty in the analysis which leads to more soundecisions.

ection 6 – swarm intelligence

Issues relating to optimised decision-making have beenddressed using swarm intelligence where Sreelaja and Pai proposen Ant Colony Optimization–Packet Filtering (ACO-PF) algorithmor packet filtering in the firewall rule-set according to the filterules. The algorithm facilitates filtering of the incoming packets innetwork by matching the rules in a rule-set. The ant agent makes

decision about the rule position in the rule-set matching with theompared field of the incoming packet based on its attractivenessowards the solution. It has been claimed that the optimised searchechnique finds the positions of the filtering rules in the rule-set

atching the compared value of the incoming packet.

ting 10 (2010) 959–962 961

In the twenty fifth paper of this Special Issue, Ghoseiri and Nad-jari present an algorithm based on multi-objective ACO to solvea bi-objective shortest path problem. It has been reported thatthe proposed ACO algorithm reduces CPU time while performancemeasures shows the solutions produced by ACO are comparablewith the label correcting solutions that take around ten timeslonger than ACO. A trade off between the CPU time and quality ofsolutions is made by implementing ACO on multi-objective shortestpath problem.

Swarm intelligence includes Particle Swarm Optimization (PSO).Raglend et al. propose an algorithm to solve the profit-based unitcommitment problem (PBUCP) under a deregulated environmentusing different PSO techniques in order to maximise the powergenerating company’s profit and schedule the generating units inthe deregulated power and reserve markets. The PSO techniquesinclude Chaotic PSO (CPSO), New PSO (NPSO) and Dispersed PSO(DPSO). Generation, spinning reserve, non-spinning reserve andsystem constraints are considered in proposed approach.

Section 7 – meta-modelling and other heuristics

The twenty-seventh paper proposes an intelligent simulationoptimisation framework for DSS by integrating meta-models inan evolutionary optimisation scheme. Within the framework, Li etal. propose a model called GA-META by integrating meta-modelsinto GA, in order to improve the efficiency and reliability of thedecision-making process. A job shop design problem is illustratedusing GA-META. A comparative study is also conducted to providegeneral recommendations on the types of meta-models for theiruse under different situations. GA is combined with the five meta-models and simulation model as well to optimise the average costof processing jobs. It has been reported that the GA-Support VectorRegression (SVR) model achieves the best solution among the otherGA-metamodels.

In the next paper Jamshidi and Karimi strive to investigate opti-misation of a nonlinear continuous multi-response problem bypresenting a two-phase hybrid meta-heuristic method, based onPareto solutions. GA is hybridised with a clustering approach andPSO algorithm in order to make a balanced relationship betweentime consuming and premature terminations. Two case studies arepresented and comparisons are conducted.

The twenty-ninth paper by Farasat et al. proposes an individual-evolutionary-based optimisation algorithm, inspired by the asexualreproduction process and known as Asexual Reproduction Optimi-sation (ARO). ARO’s adaptive search ability and its strengths andweaknesses are reported in the paper. Further, the convergence ofARO to the global optimum is analysed based on experimentationwith several frequently used benchmark functions. The perfor-mance of ARO is statistically shown to be better when comparedwith that of Particle Swarm Optimisation (PSO) on the basis ofsimulation results.

The last paper of this Special Issue reports a Parallel GreedyAlgorithm (PGA) to solve the Hybrid Flow Shop Scheduling withMultiprocessor Task (HFSMT) problem which is NP-hard. Kahra-man et al. apply PGA in two phases iteratively, viz., destruction andconstruction. Four constructive heuristic methods are proposed tosolve the HFSMT problem. A preliminary test is performed to setthe best values of the control parameters, viz., population size,subgroups number, construction method and iteration number orstopping criteria. The proposed approach is tested on a set of 240

instances adopted from the literature. The best values of the controlparameters and operators are determined by a full factorial exper-imental design using the PGA program. They report that the PGAapproach is effective in terms of reduced total completion time ormake-span for the attempted problems.

9 ompu

C

frottfwd

eilp

A

pFnfi

62 Editorial / Applied Soft C

onclusions

With these manuscripts, this Special Issue provides a platformor sharing the knowledge-base, recent research outputs and aeview of recent developments highlighting the critical aspects ofptimisation techniques in decision-making processes. It intendso attract a wider interest amongst professional engineers, prac-itioners, managers and executives so as to explore opportunitiesor improvement and to identify new trends & applications thatould enhance their organisational performance in the arena ofecision-making and optimisation methodologies.

The Guest Editors take this opportunity to gratefully acknowl-dge all the contributing authors, practising professionals fromndustry and academia, the referees and the editorial office col-eagues who have all endeavoured to bring this Special Issueossible with their latest research contributions.

cknowledgement

Dr Bhattacharya, Dr Geraghty and Dr Young acknowledge theartial support received from the Embark Initiative Postdoctoralellowship Scheme of the Irish Research Council for Science, Engi-eering and Technology (IRCSET) under the grant no. PD/2007/9 innalising this special issue.

Lead Guest EditorDr Arijit Bhattacharya ∗

The Patent Office, Mechanical Engineering Division,CP-2, Sector-V, Salt Lake City, Kolkata 700 091,

West Bengal, India

ting 10 (2010) 959–962

Guest Co-EditorsDr Sani Susanto 1

Faculty of Industrial Technology,Parahyangan Catholic University,

Jln. Ciumbuleuit 94, Bandung 40141,Indonesia

Dr John Geraghty 2

Enterprise Process Research Centre,School of Mechanical & Manufacturing Engineering,

Room No.: S-368, Dublin City University,Glasnevin, Dublin 9, Ireland

Dr Paul Young 3

Enterprise Process Research Centre,School of Mechanical & Manufacturing Engineering,

Room No.: S-374, Dublin City University,Glasnevin, Dublin 9, Ireland

∗ Corresponding author. Tel.: +91 9051638842.E-mail addresses:arijit.bhattacharya@dcu.ie,

arijit.bhattacharya2005@gmail.com(A. Bhattacharya),

sjrh@bdg.centrin.net.id (S. Susanto),john.geraghty@dcu.ie (J. Geraghty),

paul.young@dcu.ie (P. Young)

1 Tel.: +62 8164201806.2 Tel.: +353 (0)1 7007739.3 Tel.: +353 (0)1 7008216.