Safety Science 68 (2014) 169–179

Contents lists available at ScienceDirect

Safety Science

journal homepage: www.elsevier .com/locate /ssc i

A hybrid decision-making approach to measure effectiveness of safetymanagement system implementations on-board ships

http://dx.doi.org/10.1016/j.ssci.2014.04.0030925-7535/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +90 216 581 00 50; fax: +90 216 581 00 51.E-mail addresses: emreakyuz82@gmail.com, eakyuz@pirireis.edu.tr (E. Akyuz).

Emre Akyuz a,⇑, Metin Celik b

a Department of Maritime Transportation and Management Engineering, Piri Reis University, Tuzla 34940, Istanbul, Turkeyb Department of Marine Engineering, Istanbul Technical University, Tuzla 34940, Istanbul, Turkey

a r t i c l e i n f o a b s t r a c t

Article history:Received 15 August 2013Received in revised form 17 February 2014Accepted 4 April 2014

Keywords:ISM CodeSafety management systemMaritime regulationsDecision-making

Future of ship safety is recently a core topic discussed in various platforms by maritime stakeholders.Regarding this issue, it is so significant task to achieve maritime regulatory compliances with ship oper-ational requirements to ensure safe operations on-board ships. For instance, it is one of the most recentlyamendments to evaluate safety management system (SMS) effectiveness. The maritime research in thiscontext focuses on promoting a hybrid decision-making approach to measure effectiveness of safetymanagement system implementations on-board ships. The approach incorporates Analytical HierarchyProcess (AHP) and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). It determinesthe key performance indicators (KPIs) with tangible/intangible data in decision analysis which enhanceshipboard safety conditions. The main findings highlight that number of detentions, crew injuries on-board ship, and major non-conformities are considered as assessment factors of ship SMS. The proposedapproach enables to review the SMS practices systematically that is required by recent amendments ofISM Code. Thus, the proposed approach remedies the gap between safety science and maritime transpor-tation industry in terms of adopting operational data in safety analysis. Consequently, the researchoutcomes encourage the maritime researchers, safety engineers and ship operators.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Safety is the one of the key aspects of sustainable maritimetransportation. It directly deals with the management and opera-tion of ships. The achievements of the International MaritimeOrganization (IMO) related to maritime safety and marine environ-mental protection are marvellous. Principally, the IMO governs thesafety, security and environmental dimensions via regional PortState Control (PSC) authorities in accordance with the designatedmemorandum of understandings (MOUs). In current situation,the IMO declared that there are now enough regulations in placeand the problem is one of implementation and enforcement.Indeed, maritime authorities encourage the ships operators to fulfilthe requirements of adopted rules and regulation. Therefore, safetystandards on-board ships contribute to threat unsafe conditionsalong with the operation process. Hereby, it very critical aspectto ensure conformity among regulatory execution and operationalrequirements. In order to implement and enforce regulations in agood order, ship management organisation should establish an

advance monitoring system. To continuous control and verificationmight improve the maritime safety and environmental protectionstandards on a global basis (IMO, 2013).

Besides major conventions, operational safety requirements on-board ships have been supported in the form of international mar-itime codes. International safety management code (ISM Code) inthat context can be given an example in terms of safety consider-ations. The code requires establishing a safety management system(SMS) which functioning to improve safety and environmentalprevention requirement.

Since maritime safety is essential key factor in terms of mari-time transportation, several studies have been conducted overthe last decades. Tarelko (2012) explained origins of ship safetyrequirements based on the IMO policy supported with reactive orproactive actions. Furthermore, the outcomes of studies concern-ing marine accident statistics have potential to make constructivedecisions on maritime safety (Cariou et al., 2008; Mullai andPaulsson, 2011). In addition, a methodology based on fuzzy logictechnique was developed by Gaonkar et al. (2011) to evaluatesafety parameter in maritime transportation. Likewise, severaladvance models comprising Markov chains (Kolowrocki andSoszynska, 2011) and Monte Carlo simulation (Montewka et al.,2010) have been recently utilised in the same field.

170 E. Akyuz, M. Celik / Safety Science 68 (2014) 169–179

On the other hand, Heij et al. (2011) proposed a quantitativerisk assessment approach demonstrated with deficiency databasesobtained from ship inspection. In other respect, a system ofhierarchical scorecards (SHS) has been developed to evaluate theimplementation performance of maritime rules and regulations(Karahalios et al., 2011). In addition, Yang et al. (2013) reviewsthe challenges of maritime safety analysis and the differentapproaches used to quantify the risks in maritime transportation.The article has provided an update review of maritime safetyanalysis over the last decades.

The revived studies show that maritime safety is playing a crit-ical role on shipboard managements and operations. In order tomaximise maritime safety on-board ship, this study proposes ahybrid decision-making approach (AHP–TOPSIS) to assess effec-tiveness of SMS implementation on-board ship. The purposes ofthe study are highlighted as follows;

1. Enhancing safety management implementations on-boardships.

2. Developing a methodological approach to measure safetyperformance.

3. Executing ship operational procedures compliance withmaritime regulations.

In this context, this section expresses the motivation behind theresearch and literature review on maritime safety. The next onedeals with the literature review on ISM code and SMS. Then,research background upon SMS and ISM Code (2010) amendmentsis also provided. Furthermore, the methodology is proposed anddemonstrated with a case study. The final section gives the originalcontributions of the research, discussion and prospective issues forenhancement maritime safety.

2. Literature review

The ISM Code, which is enforced in 1998, refers that interna-tional safety management code for safe operation ships and pol-lution prevention. The Code was initially structured to preventmaritime accidents mainly caused from sub-standard manage-ment and operation. Regarding this issue, the purpose of Codeis to maintain international standards for safe management andoperation on-board merchant ships (IMO, 2002). The ISM Codewas first introduced in early 1980s during an investigation fortanker ships whose management standards had found inadequate(IMCO, 1982). This investigation report prompted tankership owners to consider and restructure safety policy. Moreover,it raised the awareness on-board ships and shore-basedorganisation.

The introduction of the Code affected company managementsystem significantly, beyond various shipping companies rede-signed their organisation. It required a good management practicetowards safety and pollution prevention. Therefore, an implemen-tation plan has to carry out the requirements of ISM Code con-structed by shipping company (Hunter, 1998). In detail, theshipping companies are required to develop their own policies,responsibilities and procedures under SMS. The system includesvarious procedures such as risk assessment, preventive actionplanning, maintenance planning, accident reporting, emergencyresponse plan and internal audits.

In the literature, we found a few studies mainly concern thedesign and implementations of SMS. In the first segment, Hesset al. (2011) discussed on establishing a risk assessment and clas-sification system on-board ship in accordance with ISM Code. Thepaper supported with methodology for preparation of a risk con-trol plan connected with work activities on-board ship. Anotherstudy uses a hybrid method to redesign of ISM Code procedure

to cover precautions against occupational accidents on-boardchemical tankers (Celik, 2010). The proposed method has positiveimpact to extend ISM Code procedure to OHSAS 18001:2007 stan-dards in chemical tankers operation. According to the research(Celik, 2010), the ISM Code enables a broad procedural supportto crewmembers against safety and environmental related threads.In that condition, the responsible executives ensure transformationof ISM Code procedures into operational decision supportespecially for cargo handling, tank inspection, gas freeing, tankcleaning, and tank purging operations. It is another viewpoint(Celik, 2009) to design an integrated quality and safety manage-ment system (IQSMS) for shipping operations to deal with short-falls in the shipping management. The mentioned study (Celik,2009) utilised axiomatic design principles to assess the conformitylevel of ISO integration to execution process of ISM code in mer-chant shipping. In order to provide a strategy for the safe carriageof liquid chemical cargoes in chemical tankers, a qualitativeresearch has been performed by using SWOT analysis (Arslan andEr, 2008). Another research attempting to examine the precautionpriorities during cargo operation in chemical tankers have beenstudied by Arslan (2009). In this study, AHP method is utilised toprioritize the precautions in order to explain risk assessmentoptions in chemical tanker fleet. Furthermore, another study basedon a MILP formulating has been introduced last decade (Jetlundand Karimi, 2004). In this paper, authors address the schedulingof multi-parcel chemical tankers which are carrying of numerouschemical cargoes. Moreover, a model based approach upon sys-tematic analysis has recently been introduced (Celik et al., 2013).The aim of this research is to determine the principle particularsof the optimum vessel based on the minimum construction costof chemical tankers.

A different perspective is raised by Goulielmos and Giziakis(2002) by using the fundamental of the complexity theory toreduce bureaucracy level of the SMS in practice. Likewise, anotherstudy has recently been offered to evaluate the effectiveness of theISM Code (Bhattacharya, 2012). The article reveals that there is awide gap among the perception of seafarers and company manag-ers in the ISM Code implementation. This is very challenging prob-lem subject to factual implementation of SMS on-board ships. Inaddition, Anderson (2002) conducts a wide range of survey to ana-lyse awareness of seafarers and shore-based manager upon theefficiency of ISM Code implementation on-board ship. On the effec-tiveness of ISM Code implementation, Tzannatos and Kokotos(2009) also carried out an analysis over the 268 ship accident dur-ing before and after implementation of ISM Code period. The studyreveals that implementation of the code led to decrease human-induced marine accident. Likewise, another study conducted onshipping accident on Greek-flagged ships in order to evaluate theenforcement of the ISM Code between 1995 and 2006 throughapplying of the data mining tool (Kokotas and Linardatos, 2011).

The ISM Code establishes safety management objectives and itrequires a SMS which should be established by ship managementcompany. Thereafter, the company has to set and accomplish a pol-icy for achieving those objectives. To sum up, it encourages thedevelopment of a safety culture on-board ship. The functionalrequirements of SMS associated with the operation of the shipare expressed as follows (Farthing, 1997; ISM Code, 2010); (i)safety and environmental protection policy, (ii) safety and environ-mental protection procedures, (iii) communication proceduresbetween the company and on-board ship, (iv) procedures forreporting accident or incidents, (v) emergency response plan and(vi) internal audits. The flag state of each ship has a legal right toattend management company and conduct regular audit to verifythat the company accomplish these provisions of code. If the flagstate found everything in order, they issue a certificate named doc-uments of compliance (DOC). In addition, flag state may conduct

E. Akyuz, M. Celik / Safety Science 68 (2014) 169–179 171

audit on-board ship to approve that the shore-based organisationand ships are operating in accordance with SMS. When it isachieved, flag state authority issues a certificate titled safety man-agement certificate (SMC). Each ship has to be certified with DOCand SMC for trading international seas. The certificates are validfor a period not exceeding five years from the first audit comple-tion but to be subjected to periodical audit approval (IMO, 2013).

In accordance with ISM Code clauses, each company establishestheir safety management system and appoints a designated personashore (DPA) whose responsibility is to provide a link betweenshore-based organisation and shipboard personnel. The responsi-bility of DPA should include monitoring the safety and pollutionprevention aspects of the shipboard operation (Chen, 2000). Inaddition, DPA should make sure that adequate resources andshore-based support are given to shipboard personnel. To achievethese responsibilities, DPA should be familiar with managementtools, analysis methods, etc.

After 2002 almost all of the shipping companies have estab-lished a SMS compliance with the ISM Code requirements. How-ever, it was required to adopt some amendments upon originaltext in 2008 through Maritime Safety Committee (MSC) resolution273 (85). Then, amendments were accepted on 1 January 2010 andthe latest version entered into force. Just to name a few keyamendments with new auditable requirements, the followingsare addressed; (i) assessment of all risks (article 1.2.2), (ii) measureof preventive action (article 9.2), (iii) internal audits intervals(12.1), (iv) reviewing ‘‘effectiveness of SMS’’ by company (article12.2) (ABS, 2010). For instance, in the original text related to effec-tiveness of SMS; ‘‘The Company should periodically evaluate theeffectiveness efficiency of and when needed review of the safetymanagement system in accordance with procedures establishedby the Company’’ (article 12.2) is amended. The wording ‘‘effi-ciency of and when needed review’’ is deleted. Thereby, it is clearnow that the company reviews shall be carried out periodicallyand systematically. In addition, the company shall provide evi-dence that the reviews evaluate the effectiveness of the SMS.

In the view of ISM Code (2010) amendments, the relevancemanagers and executives in shore-based organisation should takethe opportunity to review effective of SMS which means identify-ing the safety related performance, strengths, weakness and vul-nerabilities in ship fleet operations and managements.

3. Methodology

This paper presents a hybrid decision-making methodology byintegrating Analytical Hierarchy Process (AHP) and Technique forOrder Preference by Similarity to Ideal Solution (TOPSIS) in orderto measure effectiveness of SMS implementation on-board ship.The further sections introduce AHP–TOPSIS and related applicationof both methods in literature.

3.1. AHP technique

The AHP technique is first proposed by Saaty (1980) in order tosolve multiple criteria decision problems. It uses a typical pair-wise comparison technique to acquire relative weights of criteriabase upon a hierarchical structure. Briefly, AHP technique consistsof following stages (Cheng et al., 1999);

� To divide the complex problems into small part and rankthem hierarchically.

� To compare the elements by making pair-wise.� To assess the relative importance of the elements.� To unit these relevant importance and determine entire

ranking of decision alternatives.

AHP helps capture both qualitative and quantitative criteriameasurement. Thus, in the last decades, it has been widely usedfor resolving complex decision problems in numerous disciplinessuch as logistics for automobile spare parts (Li and Kuo, 2008),strategic planning for knowledge assets value creation map(Carlucci and Schiuma, 2007), knowledge management for tech-nology acquisition (Bititci et al., 2001). In addition, utilisation ofAHP technique together with hybrid method has been widelyextended in many different disciplines. For instance, Ho (2008)made researches upon application of hybrid or integrated AHPand demonstrated five works which combined with the AHP. TheAHP in SWOT analysis (Kurttila et al., 2000) is another differentexample of a combination specially developed for the purposesof practical strategic planning. Likewise, a system named SHS isdeveloped in order to assess the implementation of maritime reg-ulations and rules by combining the AHP technique with fuzzy sets(Karahalios et al., 2011). Similarly, a fuzzy TOPSIS method com-bined with AHP technique was conducted to solve the solid wastecarriage site selection problem (Onut and Soner, 2008).

Despite the popularity of the AHP, there are some disadvantagesof AHP methodology such as artificial limitation of the use of 9-scale. Another weakness of AHP is that the method may require agreat time for the pair-wise comparisons. Therefore, in order tominimise the weakness of AHP technique, it can be combined withother multicriteria decision-making tools such as TOPSIS.

This study will utilise the AHP methodology for determining thepriority weights of factors to evaluate the SMS.

3.2. TOPSIS technique

TOPSIS is a useful tool that deals with multicriteria decision-making problems. It was first introduced by Hwang and Yoon(1981). The technique helps the decision maker to organise theproblem for solving, analysing, comparing and ranking. In addition,it is a goal based approach for finding the alternative that is closestto the ideal solution. Thus, the chosen alternative should have theshortest geometric distance from the positive ideal solution andthe longest geometric distance from the negative solution. The dis-tance from positive ideal solution and negative ideal solution areconsidered simultaneously. In this method, alternatives are rankedbased on ideal solution similarity. Basically, it is considered dis-tance of options from ideal and non-ideal solution in order to mea-sure similarity of options. In the last decades, TOPSIS technique hasbeen successfully conducted to the various sectors such as manu-facturing (Milani et al., 2005), robot selection (Chu and Lin,2003), transportation (Janic, 2003), water management (Srdjevicet al., 2004), and human resource management (Chen and Tzeng,2004).

On the other hand, the hybrid AHP–TOPSIS application on mar-ine sectors is quite limited. For instance, a study on AHP–TOPSISapplication was utilised by Kandakoglu et al. (2009). In this paper,multi-methodical approach based on the application of SWOTanalysis, AHP and TOPSIS method was practiced in order to supportthe critical decision upon shipping registry selection. Anotherstudy on AHP–TOPSIS hybrid technique was performed byNooramin et al. (2012). This study takes the advantage of TOPSISand AHP hybrid technique for selecting the most efficient gantrycrane installed in marine container yard. Furthermore, AHP–TOP-SIS evaluation approach for protection of the coastal environmentin Taiwan has been recently proposed (Chang et al., 2012). In thisstudy, the authors aim is to provide an objective tool for settingcoastal protection priorities by using AHP–TOPSIS technique.

In this paper, the TOPSIS technique will be utilised to assess thesafety performance in alternative years.

172 E. Akyuz, M. Celik / Safety Science 68 (2014) 169–179

3.3. Proposed approach

In this section, a hybrid decision-making approach (AHP–TOP-SIS) will be presented to evaluate SMS effectiveness on-board ship.The AHP technique is first utilised to construct evaluation criteriahierarchy. Thereafter, a pair-wise comparison matrix is developed.Then, the criteria weights are calculated. Afterwards, TOPSISmethod is used to determine the safety performance results foralternative years. As a consequence, proposed hybrid decision-making approach (AHP–TOPSIS) acquires the benefit of both tech-niques to measure effectiveness of SMS on-board ship. A flow dia-gram for AHP–TOPSIS methodology in SMS implementationeffectiveness is illustrated in Fig. 1.

The proposed hybrid decision-making approach (AHP–TOPSIS)consists of ten steps;

Step (1) Specifying key performance indicators (KPIs): Thisincludes determination of KPIs for application. It depends on thenature of the problem, data, and expert opinion (judgement) suchas DPA and HSEQ in the decision-making process in order to estab-lish evaluation criteria of comparison matrix.

Step (2) Composing a pair-wise comparison matrix: A pair-wisecomparison matrix of criteria (A) is constructed using a scale of rel-ative importance. Saaty (1980) created a measurement 1–9 scale ofthe analytic hierarch process. Accordingly, the numbers of 1,3,5,7and 9 verbal judgements can be defined as ‘‘equal importance’’,‘‘moderate importance’’, ‘‘strong importance’’, ‘‘very strong impor-tance’’, and ‘‘extreme (absolute) importance’’. The intermediatevalues between the adjacent scale (such as 2 (weak); 4 (moderateplus); 6 (strong plus); and 8 (very, very strong)) are used for com-promise. In matrix A, each criteria aij (i,j = 1,2,3,. . .n) is the relativeimportance of ith elements compared to the jth elements. In thematrix, aij = 1 when i = j and aji = 1/aij.

A ¼

1 a12 � � � a1n

a21 1 � � � a2n

..

. ... ..

. ...

an1 an2 � � � 1

266664

377775 aii ¼ 1; aji ¼ 1=aij; aij–0 ð1Þ

Fig. 1. AHP–TOPSIS methodology in

Step (3) Calculating criterion weights (KPIs priorities) and con-sistency ratio: After composing of a pair-wise comparison matrix,normalised value of matrix is found by dividing each entry in col-umn to the sum of entries in column. Thereafter, the priorityweights of criterion are calculated. The average of value in eachrow gives estimate of relative weights of criterion. The normalisa-tion of matrix and priority weights of criterions (W1, W2,. . ., Wj) canbe calculated with following equations;

rij ¼aijPni¼1aij

; i ¼ 1;2; . . . ;n and j ¼ 1;2; . . . ;n ð2Þ

Wj ¼1n�Xn

i¼1

aij; i ¼ 1;2; . . . ;n and j ¼ 1;2; . . . ;n ð3Þ

In order to provide consistency of data provided in methodol-ogy, Saaty proposed an equation to verify whether the matrix isconsistent or not. Accordingly, consistency index (CI) can be calcu-lated as follows (4);

CI ¼ kmax: � nn� 1

ð4Þ

In equation, n is the order of the matrix, and kmax is maximumeigenvalue of the matrix and it can be found with following equa-tion (Vargas, 1982).

Xn

j¼1

aijwj ¼ kmax:wi ð5Þ

In order to determine reasonable consistency, a consistencyratio (CR) value should be calculated. If the CR value is found equalor less than 0.10, the judgments are considered as consistent. Theformulation of CR can be stated as follows;

CR ¼ CI=RI ð6Þ

The Random index (RI) value is illustrated on the Table 1. The RIis the indicator for random and it is subjected to the number ofitems that is compared in matrix (Saaty, 1994).

SMS effectiveness evaluation.

Table 1The values of random index.

n 1 2 3 4 5 6 7 8 9 10RI 0 0 0.58 0.90 1.12 1.24 1.32 1.41 1.45 1.49

Bold values show the polarity of attributes.

Table 2KPIs description, code and sources table.

Name of KPIs CodeofKPIs

Sources

Number of deficiency observed on-boardships (�)

KPI1 Knudsen and Hassler(2011)

Number of completed training on-boardships (+)

KPI2 ISM Code (2010)

Number of major non-conformity observedon-board ships (�)

KPI3 DNV (2012)

Number of detention (�) KPI4 Cariou et al. (2009)Number of near-miss reported by ships (�) KPI5 Storgard et al. (2012)Number of successful psychometric test

applied for officer (+)KPI6 Ek and Olsson (2000)

Number of crew injury observed on-boardships (�)

KPI7 Storgard et al. (2012)

DPA internal audit judgement (+) KPI8 Management companyHSEQ Manager audit judgement (+) KPI9 Management company

E. Akyuz, M. Celik / Safety Science 68 (2014) 169–179 173

Step (4) Constructing decision matrix (D): This step is to repre-sent all information available for the attribute in the decisionmatrix. The structure of the decision matrix can be defined asfollows;

D ¼

c1 c2 c3 � � � cn

A1 x11 x12 x13 � � � x1n

A2 x21 x22 x23 � � � x2n

A3 x31 x32 x33 � � � x3n

..

. ... ..

. ... . .

. ...

Am xm1 xm2 xm3 � � � xmn

26666666664

37777777775

ð7Þ

where Ai = ith alternative related and xij is the performance value ofalternative with respect to criterion cj.

Step (5) Calculating normalised decision matrix: This step nor-malizes the decision matrix D by using the following formula;

rij ¼xijffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiPm

i¼1x2ij

q ; i ¼ 1;2;3; . . . ;m and j ¼ 1;2;3; . . . ;n: ð8Þ

Step (6) Calculating weighted normalised decision matrix: Inorder to construct weighted normalised decision matrix (vij), asso-ciated weight is to be multiplied with its normalised decisionmatrix. The calculation is as follows;

v ij ¼ wj � rij; i ¼ 1;2; . . . ; n; j ¼ 1;2; . . . ; n ð9Þ

where wj is the weight of the jth attribute or criterion.Step (7) Determining the positive ideal solution (PIS) and nega-

tive ideal solution (NIS): The PIS and NIS values can be determinedvia taking the maximum and minimum values within the row ofweighted normalised decision matrix.

Aþ ¼ fðmax v ijjj 2 JÞ or ðmin v ijjj 2 J0Þ for i ¼ 1;2; . . . ;mg¼ fvþ1 ;vþ2 ; . . . ;vþn g ð10Þ

A� ¼ fðmin v ijjj 2 JÞ or ðmaxv ijjj 2 J0Þ for i ¼ 1;2; . . . ;mg¼ fv�1 :;v�2 ; . . . ; v�n g

where J = 1,2,3,...,n. is associated with benefit (positive criteria) andJ0 = 1,2,3,...,n is associated with cost (negative criteria).

Step (8) Calculating of separation measure: The separation ofeach alternative from the PIS can be found by following equations;

Sþi ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiXn

j¼1ðv ij � vþj Þ

2r

; i ¼ 1;2; . . . ;m ð11Þ

Likewise, the separation from the NIS can be defined as;

S�i ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiXn

j¼1ðv ij � v�j Þ

2r

; i ¼ 1;2; . . . ;m ð12Þ

Step (9) Calculating the relative closeness to the ideal solution:The relative closeness has been measured by following equation;

C�i ¼S�i

Sþi þ S�i; 0hC�i h1; i ¼ 1;2; . . . ;m ð13Þ

Step (10) Ranking the preference order (SMS effectiveness eval-uation): The final step is provided to rank alternatives in accor-dance with the descending order of C�i . This step provides acomparison of alternative years in SMS effectiveness.

3.4. KPIs for SMS

Key performance indicators (KPIs) are generally used to mea-sure progress or to monitor trends which can be used to demon-strate where further improvements or resources are required. Inorder to monitor performance of the SMS implementation, relevantrecords and evidences which are gathered from the ship can be uti-lised. So that continuous monitoring of the KPIs will give ideaabout the state of SMS implementations on-board ship. If therewould be any shortfalls on SMS design, it should be revised toincrease its utility. Moreover, the KPIs based analysis improvessafety performance on-board ships. In addition, application of KPIsupon SMS implementation will provide to systematic review onsafety conditions.

The determined KPIs will help to measure effectiveness of SMSimplementations on-board ships by using the hybrid decision-making approach (AHP–TOPSIS). The KPIs are provided on theTable 2.

3.4.1. Number of deficiency observed on-board ship (KPI1)Deficiency generally refers lack of safety and environment

requirements on-board ship. It deals with the non-fulfilment ofregulatory requirements. The main categories of shipboard defi-ciencies, observed by PSC authorities, are given as follows;

Certificates and documentation (i.e. minimum safe manningdocument)

� Structural condition (i.e. closing devices/watertight doors).� Water/weather tight condition (i.e. railing, gangway, walk-

way and means for safe passage).� Life saving appliances (i.e. lifeboat inventory).� Pollution prevention (i.e. control of discharge of oil).

The deficiencies should be identified and recorded. Hence,records on the deficiencies per year are one of the key indicatorsto measure effectiveness of SMS implementations on-board ships.

3.4.2. Number of completed training on-board ship (KPI2)The purpose of training on board is to provide improvement

crew awareness on safety and environment requirements. Theshore-based managers dispatch the training requirements to shipauthority to improve competency requirements. Therefore, num-ber of completed training on-board ship in a year is consideredessential indicator for processing of SMS on-board ship.

174 E. Akyuz, M. Celik / Safety Science 68 (2014) 169–179

3.4.3. Number of major non-conformity (KPI3)Major non-conformity is defined as a serious threat which may

result in vital failure to safety of crew, ship or environment thatrequires immediate corrective actions (IMO, 2013). In operationallevel, PSC officers conduct a survey on-board ship. If PSC officerfinds major non-conformities during inspection, it must be recti-fied before vessel departure. For instance, lack of ISM certificate,unfamiliar new crew with their duties, lack of communication pro-cedure in emergency situation, etc. Since the major non-conformi-ties are critical records, it will help to measure effectiveness of SMSimplementations on-board ships.

3.4.4. Number of detention (KPI4)If the corrective/preventive action procedures are not remedy

the major non-conformities, the ship is not allowed to sail in inter-national waters. This is known as detention. Since a well-designed/implemented SMS will prevent such kind of events, the number ofdetention is another critical indicator for effectiveness of SMSimplementations.

3.4.5. Number of near-miss (KPI5)It is defined as unexpected event which have not resulted in

loss of life or injury but had to potential to do. Near-misses arerecorded on-board ship to make causation analysis in order to pre-vent reoccurrence. Hence, number of near-miss incidents in a yearcan provide an opportunity to evaluate effectiveness of SMS imple-mentation on-board ship.

3.4.6. Number of successful psychometric test applied for officer (KPI6)According to maritime regulations, each company should pro-

vide to ship with qualified and medically fit seafarers. In order tofulfil this requirement, shore-based managers have recentlyapplied a psychometric test for officer. The test provides an objec-tive way to monitor seafarer’s physical and mental performance.The record of psychometric test is kept in company and on-boardship. Accordingly, the number of successful psychometric testresults per year can be utilised as human factor parameter toassess effectiveness of SMS implementation on-board ship.

3.4.7. Number of crew injury on-board ship (KPI7)Crew injuries are very common issue on-board ship and ship

crew always face a high risk at sea conditions. As clearly pointout in ISM Code, ensuring safety at sea, prevention of human injuryor loss of life is clarified as one of the main objectives. Therefore,number of crew injuries on-board ship must keep a record andbeing reported to shore-based organisation. The crew injury rateis an indicator that shows whether SMS implementation on-boardis in a good order or not.

3.4.8. DPA internal audit judgement (KPI8)The responsibility of DPA is clearly defined in ISM Code as mon-

itoring the safety and pollution prevention aspect of the ship. DPAattends on-board ship to conduct an internal audit regularly inorder to ensure that good SMS practice on-board ships. Thus, thegeneral review of DPA can be adopted into SMS performanceassessment. The verbal judgement based on a 1–5 rating scale(Belz and Kow, 2010) provides incorporation of an importantparameter into SMS performance review. Furthermore, rating scalecan include followings; ‘‘1’’ refers poor performance, ‘‘2’’ indicatesfair performance, ‘‘3’’ refers good performance, ‘‘4’’ refers very goodperformance and ‘‘5’’ indicates excellent performance for the year.

3.4.9. HSEQ Manager audit judgement (KPI9)In ship management companies, health, safety, environment and

quality (HSEQ) department has recently been established toimprove the safety, quality and environment performance in ship

management and operation. The HSEQ department mainly concen-trates on adopting quality principles into health, safety and environ-ment considerations. The HSEQ manager judgements are taking intoaccount to gain a different viewpoint in SMS effectiveness analysis.The verbal judgement, based on a 1–5 rating scale, is used.

4. Application

In this section the proposed AHP–TOPSIS approach will be uti-lised to measure effectiveness of SMS implementations on-boardships. To demonstrate the model, it is contacted with a prestigiousshipping company which has a chemical tanker fleet. The fleet ageis ranging between 2 to 15 years old and ships size varying 3000–17,000 deadweight tons.

4.1. Analysis of respondents

The data contains qualitative and quantitative information forKPIs basis per year. The quantitative data (KPI1 to KPI7) were pro-vided through data records which are consisting of PSC reports,vetting control reports and company internal audit reports. Thedata received from company is available for the last three years.In addition, subjective data has received from maritime experts.The expert profile contains professional managers (DPA and HSEQdepartment) and marine superintendents who have seagoing back-ground and professional execution experiences. While constitutingthe expert group, average seagoing experiences (six years) andshore-based management experiences (five years) are taken intoaccount. Hence, the judgements of the experts are considered ina group consensus via a brainstorming meeting. In this meeting,technical and operational aspects of the problem are consideredin the evaluation of safety related activities on-board. The groupprovide data/judgements for pairwise comparison of KPIs and thesubjective data for KPI8 and KPI9 in the content of decision matrix.

4.2. Data collection

Considering the importance of ISM Code implementation on-board ship, this study is conducted by using real data from opera-tional level. It measures the effectiveness of SMS implementationon-board ship by using the KPIs. Accordingly, the shore-basedorganisation can realise whether SMS implementation on-boardis useful or not.

In this study, a shipping company operating chemical tankerfleet is decided. Execution of chemical tanker fleet is requiredadvance management system due to the critical operationalaspects. In this demonstration, the data were gathered by review-ing company SMS records, related documents and correspondencecommunication with ships. In addition, PSC reports and vettingreports were benefitted to analyse deficiency, major non-confor-mity and detention. On the other hand, linguistic data has receivedfrom DPA and HSEQ department of the company for KPI8 and KPI9

subjective judgement.

4.3. Empirical analysis

In application step, KPIs assist to measure effectiveness of SMSimplementation on-board ship. These indicators can also be helpfulto review performance of SMS implantation on-board ships. More-over, knowing how to prioritize KPIs can help us streamline thedecision-making process of application.

After determining the KPIs for SMS effectiveness application, apair-wise comparison matrix is established in accordance with1–9 scale of the analytic hierarchy process which is showing theintensity of importance each criterion. The judgements on KPIs,

Table 3DPA judgements on KPIs.

KPI1 KPI2 KPI3 KPI4 KPI5 KPI6 KPI7 KPI8 KPI9

KPI1 Equalimportance

Strongimportance

1/Moderateplus

1/Moderateimportance

Strongimportance

Very strongimportance

1/Strongimportance

1/Moderate plus 1/Moderateimportance

KPI2 Reciprocal Equalimportance

1/Strongplus

1/Very strongimportance

Moderateplus

Strong plus 1/Moderateimportance

Weak ModerateImportance

KPI3 Equalimportance

1/Moderate plus Strongimportance

Strongimportance

1/Moderateimportance

Moderateimportance

Strong plus

KPI4 Equal importance Strong plus Extremeimportance

Weak Strongimportance

Very, verystrong

KPI5 Equalimportance

ModerateImportance

1/Strongimportance

1/Moderateimportance

Moderateimportance

KPI6 Equalimportance

1/Very strongimportance

1/Moderateimportance

Moderateimportance

KPI7 Equal importance ModerateImportance

Strongimportance

KPI8 Equalimportance

Moderateimportance

KPI9 Equalimportance

E. Akyuz, M. Celik / Safety Science 68 (2014) 169–179 175

which are given in Table 3, were received from the company DPAto construct a pair-wise comparison matrix

The corresponding numerical equivalents for each judgementare shown in Table 4. For example, ‘‘number of completed trainingon-board ship’’ (KPI2) has weak importance than ‘‘DPA internalaudit judgement’’ (KPI8); therefore, number 2 is assigned for thiscomparison. Likewise, the reciprocal equation of KPI8 to KPI2 isassigned 1/2 as proposed in Eq. (1).

After composing of a pair-wise comparison matrix, the valuesare need to normalised basis Eq. (2). It is found by dividing eachentry in column to the sum of entries in column. In Table 5, norma-lised value of each KPI is illustrated.

The priority weights of KPIs are calculated in accordance withEq. (3). The average of value in each row gives estimate of relativeweights of KPIs. Numerical weight values and percentages of eachKPI are provided in Table 6.

According to the Table 6, KPI4 (number of detention) has thehighest weight criterion (0.31) and it is percentage is 30.94% inoverall. Thereafter, KPI7 (number of crew injury on-board ship)and KPI3 (number of major non-conformity) are ranking respec-tively. Since the KPI4 is ranked on the top of priority weight table,

Table 4Pair-wise comparison matrix.

KPI1 KPI2 KPI3 KPI4 KPI5 KPI6 KPI7 KPI8 KPI9

KPI1 1.00 5.00 1/4 1/3 5.00 7.00 1/5 1/4 1/3KPI2 1/5 1.00 1/6 1/7 3.00 6.00 1/3 2.00 3.00KPI3 4.00 6.00 1.00 1/4 5.00 5.00 1/3 3.00 5.00KPI4 3.00 7.00 4.00 1.00 7.00 9.00 3.00 5.00 7.00KPI5 1/5 1/3 1/5 1/7 1.00 3.00 1/5 1/3 3.00KPI6 1/7 1/6 1/5 1/9 1/3 1.00 1/7 1/3 3.00KPI7 5.00 3.00 3.00 1/3 5.00 7.00 1.00 3.00 5.00KPI8 4.00 1/2 1/3 1/5 3.00 3.00 1/3 1.00 3.00KPI9 3.00 1/3 1/5 1/7 1/3 1/3 1/5 1/3 1.00

Table 5Normalised pair-wise comparison matrix.

KPI1 KPI2 KPI3 KPI4 KPI5 KPI6 KPI7 KPI8 KPI9

KPI1 0.05 0.21 0.03 0.13 0.17 0.17 0.03 0.02 0.01KPI2 0.01 0.04 0.02 0.05 0.10 0.15 0.06 0.13 0.10KPI3 0.19 0.26 0.11 0.09 0.17 0.12 0.06 0.20 0.16KPI4 0.15 0.30 0.43 0.38 0.24 0.22 0.52 0.33 0.23KPI5 0.01 0.01 0.02 0.05 0.03 0.07 0.03 0.02 0.10KPI6 0.01 0.01 0.02 0.04 0.01 0.02 0.02 0.02 0.10KPI7 0.24 0.13 0.32 0.13 0.17 0.17 0.17 0.20 0.16KPI8 0.19 0.02 0.04 0.08 0.10 0.07 0.06 0.07 0.10KPI9 0.15 0.01 0.02 0.05 0.01 0.01 0.03 0.02 0.03

it is considered as the most critical factor in terms of safety man-agement system implementation on-board ship. Subject to thevariety and number of detention, the ship might not be allowedto sail by port authorities until the necessary rectification fordetentions have been completed by master of vessel or shore-based management company. Moreover, number of detentionslead to lose prestige for shipping company. Therefore, a well organ-ised SMS is designed and implemented by shore-based manage-ment company to prevent possible detention in advance.

The KPI7 is the second most crucial shortfall factor in accor-dance with effectiveness of safety management system implemen-tations on-board ships. Since ship crew face a high risk at seaconditions, crew injury rate gives an idea whether effectivenessof SMS implementation on-board is properly fulfilled or not.Thereby, as the second important factor, the SMS implementationshall be organised to prevent crew injury on-board ship.

The KPI3 has the third highest weight criterion and shortfalleffectiveness of safety management system implementation on-board ships. The major non-conformity shall be prevented byimplementation SMS on-board ship properly. In case major non-conformities do not rectify on-time, the ship will face detain inport.

After calculating priority criterion weight, the consistencydegree of the matrix is controlled to satisfy the consistency of judg-ments in the pair-wise comparison. The consistency ratio (CR) ofmatrix can be calculated by using Eqs. (4)–(6). The random index(RI) value will be 1.45 since nine factors are compared in matrix.Consequently, the CR value can be found as 0.093. Since the CRvalue is less than 0.10, all data inserted in comparison matrix isconsidered as consistency.

In the next step, decision matrix is established basis Eq. (7).Table 7 is illustrating data records and judgements provided bycompany on a yearly basis. The data demonstrates the total num-bers that have occurred in that year for fleet. For instance, totallythirty-four deficiencies have been observed in 2010. In addition,two near-miss events have been reported in 2011. Beside numeri-cal data information, verbal judgement of DPA and HSQE depart-ments take a place in table basis 1–5 rating scale. For example,DPA assigned with FP (fair performance) for 2010 years in termsof reviewing the safety and environmental related performanceof ships. On the other hand, same year has been evaluated as GP(good performance) by the HSQE manager in accordance with theresponsibility scope.

The data received from DPA are utilised to compose initial deci-sion matrix in Table 8 where negative factors (cost attributes) arereciprocally inserted. For instance, KPI7 (number of crew injury

Table 6KPIs priorities.

Criterion Weight Percentage (%)

KPI1 0.09 9.06KPI2 0.07 7.32KPI3 0.15 15.13KPI4 0.31 30.94KPI5 0.04 4.01KPI6 0.03 2.87KPI7 0.19 18.80KPI8 0.08 8.04KPI9 0.04 3.83

Table 7Data on KPIs.

2010 2011 2012

KPI1 34 37 26KPI2 21 17 14KPI3 8 4 5KPI4 2 1 1KPI5 5 2 3KPI6 57 52 59KPI7 4 2 2KPI8 FP VGP GPKPI9 GP VGP VGP

Table 8Initial decision matrix.

Scale Weight 2010 2011 2012

KPI1 Numbers 0.09 1/34 1/37 1/26KPI2 Numbers 0.08 21 17 14KPI3 Numbers 0.15 1/8 1/4 1/5KPI4 Numbers 0.29 1/2 1 1KPI5 Numbers 0.04 1/5 1/2 1/3KPI6 Numbers 0.03 57 52 59KPI7 Numbers 0.20 1/4 1/2 1/2KPI8 5-Scale judgement 0.08 FP VGP GPKPI9 5-Scale judgement 0.04 GP VGP VGP

Table 9Decision matrix.

Weight 2010 2011 2012

KPI1 0.09 0.03 0.03 0.04KPI2 0.08 21.00 17.00 14.00KPI3 0.15 0.13 0.25 0.20KPI4 0.29 0.50 1.00 1.00KPI5 0.04 0.20 0.50 0.33KPI6 0.03 57.00 52.00 59.00KPI7 0.20 0.25 0.50 0.50KPI8 0.08 2.00 4.00 3.00KPI9 0.04 3.00 4.00 4.00

Table 10Normalised decision matrix.

Weight 2010 2011 2012

KPI1 0.09 0.53 0.49 0.69KPI2 0.08 0.69 0.56 0.46KPI3 0.15 0.36 0.73 0.58KPI4 0.29 0.33 0.67 0.67KPI5 0.04 0.32 0.79 0.53KPI6 0.03 0.59 0.54 0.61KPI7 0.20 0.33 0.67 0.67KPI8 0.08 0.37 0.74 0.56KPI9 0.04 0.60 0.80 0.80

Table 11Weighted normalised decision matrix.

2010 2011 2012

KPI1 0.05 0.04 0.06KPI2 0.05 0.04 0.04KPI3 0.06 0.11 0.09KPI4 0.10 0.20 0.20KPI5 0.01 0.03 0.02KPI6 0.02 0.02 0.02KPI7 0.07 0.13 0.13KPI8 0.03 0.06 0.04KPI9 0.02 0.03 0.03

Table 12Positive/negative ideal solution.

Max Min

KPI1 0.06 0.04KPI2 0.05 0.04KPI3 0.11 0.06KPI4 0.20 0.10KPI5 0.03 0.01KPI6 0.02 0.02KPI7 0.13 0.07KPI8 0.06 0.03KPI9 0.03 0.02

Table 13Distance calculation, relative closeness and ranking.

2010 2011 2012

S+ 0.14 0.02 0.03S� 0.02 0.14 0.13C⁄ 0.12 0.86 0.79Rank 3 1 2

Fig. 2. SMS effectiveness evaluation results.

Table 14KPI based distances to PIS.

2010 2011 2012

KPI1 0.015 0.019 0.000KPI2 0.000 0.010 0.018KPI3 0.056 0.000 0.022KPI4 0.098 0.000 0.000KPI5 0.019 0.000 0.011KPI6 0.001 0.002 0.000KPI7 0.066 0.000 0.000KPI8 0.030 0.000 0.015KPI9 0.007 0.000 0.000

176 E. Akyuz, M. Celik / Safety Science 68 (2014) 169–179

E. Akyuz, M. Celik / Safety Science 68 (2014) 169–179 177

on-board ship) has been reported as four times in 2010. Since num-ber of crew injury has negative effect on SMS effectiveness evalua-tion, it should be inserted as reciprocal. Therefore, KPI7 value(number of crew injury on-board ship) for 2010 is inserted as 1/4in decision matrix.

The equivalent values of data, called as decision matrix, areillustrated in Table 9. In this section, verbal judgement of DPAand HSQE manager are converted to numerical value in accordancewith 1–5 rating scale. For example, DPA assigned a judgement forfleet conditions in 2012 are good (GP). Therefore, it is converted to3 in accordance with 1–5 rating scale.

The decision matrix is normalised by using the Eq. (8). The nor-malised decision matrix is shown in Table 10.

Thereafter, weighted normalised decision matrix is calculatedin accordance with Eq. (9). The results are illustrated in Table 11.

Fig. 3. SMS overview bas

Fig. 4. SMS overview bas

Positive ideal solution (best) and negative ideal solution (worst)is determined by using Eq. (12). Table 12 provides PIS and NIS val-ues for each KPI.

After doing a separation measurement by using Eqs. (11) and(12) for PIS and NIS, the relative closeness of criterion is calculatedusing the Eq. (12). The results are shown in Table 13.

Finally, the criterion are arranged in descending order in accor-dance with their relative closeness and ranked the preferenceorder.

4.4. Findings

According to the results, the SMS effectiveness in 2011 reachesthe highest value when compare the previous year. Fig. 2illustrates the SMS effectiveness evaluation results in the period

ed on KPIs in 2010.

ed on KPIs in 2011.

Fig. 5. SMS overview based on KPIs in 2012.

178 E. Akyuz, M. Celik / Safety Science 68 (2014) 169–179

of 2010–2012. It seems that there is a significant raise in 2011 interms of SMS implementation benefits. However, the system eval-uation reveals a slight decrease in 2012.

The distance values of each KPI for alternative years to PIS cangive idea to decision makers (ship operators and managers) aboutcritical issues in SMS implementations. The values, KPI based dis-tances to PIS, are provided in Table 14.

For cluster based analysis, Figs. 3–5 illustrate the vulnerabilitiesin the system through 2010, 2011 and 2012 respectively.

As it is seen in Fig. 3, KPI4 (number of detention), the distalpoint, is determined as the most remarkable factor to take intoconsideration. On the other hand, KPI7 (number of crew injuryobserved on-board ships) and KPI3 (number of major non-confor-mity observed on-board ships) are the other critical aspects in2010.

In 2011, KPI1 (number of deficiency found on-board ships) andKPI2 (number of completed training on-board ships) are the mostcritical aspects according to Fig. 4. As the most successful yearwithin alternatives, the variety of the KPIs cumulates in the centre.

Fig. 5 points out that KPI3 (number of major non-conformityobserved on-board ships), KPI2 (number of completed trainingon-board ships), KPI5 (number of near-miss reported by ships)and KPI8 (DPA internal audit judgement) are the key aspects inorder to increase effectiveness of SMS implementation.

5. Conclusion and discussion

Regulatory compliance matter is an onerous task recentlydiscussed various platforms in maritime society. This problemrequires establishing robust models in order to make sensitivedecision analysis on regulation requirements and operationalconditions. For instance, the latest amendment to ISM Code neces-sitates performing an effectiveness analysis to SMS implementa-tions on-board ships. Solution of compliance matter in ISM Codeespecially deals with reducing bureaucracy, integrated documen-tation management, effecting reporting, continuous improvementand active monitoring with performance indicators. At this point,the managers in shore-based organisation should develop a safetymonitoring system that complies with the auditable requirementsof international authorities.

This research prompted a hybrid-decision model to monitorthe implementation performance of SMS via monitoring the KPIsdata. The model is based on AHP and TOPSIS technique to prior-itize and use KPIs data under unique frame on which safety per-formance is measured. Since the evaluation of safety is verychallenging process, the proposed approach has managed to fillthe gap both in safety science literature theoretically and mari-time industry practically. The main findings of the research showthat number of detentions, crew injuries on-board ship, andmajor non-conformities are the key indicator to make decisionon ship SMS integration. Hence, the proposed approach enablesto review the SMS practices systematically that is required byrecent amendments of ISM Code. Therefore, original contributionsof the research outcomes to maritime researchers, safety engi-neers and ship operators can be highly appreciated. In addition,the paper is expected to contribute on-going efforts towardsimprovement of chemical tanker SMS since chemical tanker oper-ations require a high level of safety and environmental-relatedprecautions when compared with the other types of merchantships. The proposed model provides to analyse the SMS practicein chemical tanker fleet safety management by evaluating num-ber of detention, crew injuries and major non-conformities. Thus,monitoring the implementation performance of SMS should becarried out in order to reduce catastrophic factor that may affectchemical tanker safety. Moreover, management of chemical tan-ker shipping company can utilise the proposed approach to avoidany detention or crew injuries or major non-conformities inadvance by measuring effectiveness of SMS implementationson-board chemical tankers. Thereby, the vessel is not arrestedor detained by PSC officer (Cariou et al., 2009). The followingpoints of the research can briefly be highlighted;

(i) The proposed model can utilise both quantitative/qualitativedata in safety analysis.

(ii) It supports the required solutions to ISM Code (2010)amendments.

(iii) The model can be adopted into SMS documentation.(iv) The research encourages the safety practitioners and mari-

time executives to establish model base system in ship oper-ation and management.

E. Akyuz, M. Celik / Safety Science 68 (2014) 169–179 179

Furthermore, the proposed approach can be modified to involvelong range data with variety of KPIs. In addition, the model can besupported with information technologies to enhance implementa-tion practice.

References

ABS, 2010. ISM Code Update. Amendments to the ISM Code MSC 85 Resolution, vol.273(85).

Anderson, P. 2002. Managing safety at sea. D. Phil thesis, UK: Middlesex University.Arslan, O., Er, I.D., 2008. SWOT analysis for safer carriage of bulk liquid chemicals in

tankers. J. Hazard. Mater. 154, 901–913.Arslan, O., 2009. Quantitative evaluation of precautions on chemical tanker

operations. Process Saf. Environ. Prot. 87, 113–120.Belz, A., Kow, E., 2010. Comparing rating scales and preference judgements in

language evaluation. In: Proceedings of the 6th International Natural LanguageGeneration Conference, pp 07–15.

Bhattacharya, S., 2012. The effectiveness of the ISM Code: a qualitative enquiry.Mar. Policy 36, 528–535.

Bititci, U.S., Suwignjo, P., Carrie, A.S., 2001. Strategy management throughquantitative modelling of performance measurement systems. Int. J. Prod.Econ. 69 (1), 15–22.

Cariou, P., Mejia, M.Q., Wolff, F.C., 2009. Evidence on target factors used for portstate control inspections. Mar. Policy 33, 847–859.

Cariou, P., Mejia, M.Q., Wolff, F.C., 2008. On the effectiveness of port state controlinspections. Transport. Res. Part E – Logis. Transpiration 44 (3), 491–503.

Carlucci, D., Schiuma, G., 2007. Knowledge assets value creation map – assessingknowledge assets value drivers using AHP. Exp. Syst. Appl. 32 (3), 814–821.

Celik, F., Erturk, I., Turan, E., 2013. Investigation of main particulars subject tominimum building cost for chemical tankers. Ocean Eng. 73, 32–37.

Celik, M., 2010. Enhancement of occupational health and safety requirements inchemical tanker operations: the case of cargo explosion. Saf. Sci. 48, 195–203.

Celik, M., 2009. Designing of integrated quality and safety management system(IQSMS) for shipping operations. Saf. Sci. 47, 569–577.

Chang, H.K., Liou, J.C., Chen, W.W., 2012. Protection priority in the coastalenvironment using a hybrid AHP-TOPSIS method on the Miaoli Coast, Taiwan.J. Coast. Res. 28 (2), 369–374.

Chen, L., 2000. Legal and practical consequences of not complying with ISM code.Maritime Policy Manag: Flagship J. Int. Shipping Port Res. 27 (3), 219–230.

Chen, M.F., Tzeng, G.H., 2004. Combining gray relation and TOPSIS concepts forselecting an expatriate host country. Math. Comput. Model. 40, 1473–1490.

Cheng, A.C., Yang, B.K., Hwang, C., 1999. Evaluating attack helicopters by the AHPbased on linguistic variable weight. Eur. J. Oper. Res. 116 (2), 423–435.

Chu, T.C., Lin, Y.C., 2003. A fuzzy topsis method for robot selection. Int. J. Adv.Manuf. Technol. 21, 284–290.

DNV, 2012. Port state control-synopsis of frequent findings and detention items.DNV serving the maritime industry.

Ek, A., Olsson, U.M.M., 2000. Safety culture onboard ships. In: Proceedings of theHuman Factors and Ergonomics Society Annual Meeting. vol. 44, no. 27, pp.320–322.

Farthing, B. 1997. The ISM Code: A leap forward or a return to old standards. Balticand International Maritime Council (BIMCO). Review. London, pp. 93–95.

Gaonkar, R.S.P., Xie, M., Ng, M.M., Habibullah, M.S., 2011. Subjective operationalreliability assessment of maritime transportation system. Exp. Syst. Appl. 38(11), 13835–13846.

Goulielmos, A.M., Giziakis, C.B., 2002. Marine accident prevention: an evaluation ofthe ISM code by the fundamentals of the complexity theory. Disaster PreventionManag. 11 (1), 18–32.

Heij, C., Bijwaard, G.E., Knapp, S., 2011. Ship inspection strategies: Effects onmaritime safety and environmental protection. Transport. Res. Part D-TransportEnviron. 16 (1), 42–48.

Hess, M., Kos, S., Njegovan, M., 2011. Assessment and control of operational risks onboard ships in accordance with the ISM Code. J. Maritime Stud. 25 (2), 405–416.

Ho, W., 2008. Integrated analytic hierarchy process and its applications, a literaturereview,. Eur. J. Operational Res. 186 (1), 211–228.

Hunter, J.A.D., 1998. Shipowner’s perspective of its development: A success or afailure?. New Safety Culture Conference, Institute of Marine Engineers, pp. 3–7.

Hwang, C.L., Yoon, K.P., 1981. Multiple Attribute Decision Making: Methods andApplications. Springer-Verlag, New York.

Janic, M., 2003. Multicriteria evaluation of high-speed rail, Transrapid Maglev, andair passenger transport in Europe. Transport. Plann. Technol. 26, 491–512.

Jetlund, A.S., Karimi, I.A., 2004. Improving the logistics of multi-compartmentchemical tankers. Comput. Chem. Eng. 28, 1267–1283.

IMCO, 1982. Tanker casualty investigations: Report of the tanker accident workinggroup by ICS, OCIMF and INTERTANKO. Presented at maritime safety committee46th session agenda no.18. (MSC46/18/726February1982).

IMO, 2013. ISM code and guidelines on implementation of the ISM code. <http://www.imo.org>.

IMO, 2002. International Safety Management Code and Revised Guide Lines onImplementation of the ISM Code by Administrations. IMO, London, 2002.

ISM Code, 2010. ISM Code and Guidelines on Implementation of the ISM Code 2010.Resolution A.741(18) as amended by MSC.104(73), MSC.179(79), MSC.195(80)and MSC.273(85). London, UK.

Kandakoglu, A., Celik, M., Akgun, I., 2009. A multi-methodological approach forshipping registry selection in maritime transportation industry. Math. Comput.Model. 49, 586–597.

Karahalios, H., Yang, Z.L., Williams, V., Wang, J., 2011. A proposed system ofhierarchical scorecards to facilitate the implementation of maritimeregulations. Saf. Sci. 49 (3), 450–462.

Kokotas, X.D., Linardatos, S.D., 2011. An application of data mining tools for thestudy of shipping safety in restricted waters. Saf. Sci. 49, 192–197.

Kolowrocki, K., Soszynska, J., 2011. On safety analysis of complex technicalmaritime transportation systems. J. Risk Reliab. 225 (3), 345–354.

Knudsen, O.F., Hassler, B., 2011. IMO legislation and its implementation: Accidentrisk, vessel deficiencies and national administrative practices. Mar. Policy 35,201–207.

Kurttila, M., Pesonen, M., Kangas, J., Kajanus, M., 2000. Utilizing AHP in SWOTanalysis: a hybrid method and its application. Forest Policy Econom. 1, 41–45.

Li, S.G., Kuo, X., 2008. The inventory management system for automobile spare partsin a central warehouse. Exp. Syst. Appl. 34 (2), 1144–1153.

Milani, A.S., Shanian, A., Madoliat, R., 2005. The effect of normalization norms inmultiple attribute decision making models: a case study in gear materialselection. Struct. Multidiscip. Optimization 29, 312–318.

Montewka, J., Hinz, T., Kujala, P., Matusiak, J., 2010. Probability modelling of vesselcollisions. Reliab. Eng. Syst. Saf. 95 (5), 573–589.

Mullai, A., Paulsson, U., 2011. A grounded theory model for analysis of marineaccidents. Accid. Anal. Prev. 43 (4), 1590–1603.

Nooramin, A.S., Sayareh, J., Moghadam, M.K., Alizmini, H.R., 2012. TOPSIS and AHPtechniques for selecting the most efficient marine container yard gantry crane.Oper. Res. Decis. Theory 49 (2), 116–132.

Onut, S., Soner, S., 2008. Transshipment site selection using the AHP and TOPSISapproaches under fuzzy environment. Waste Manage. (Oxford) 28, 1552–1559.

Saaty, T.L., 1994. How to make a decision: the analytic hierarchy process. Interfaces24 (6), 19–43.

Saaty, T.L., 1980. The Analytic Hierarchy Process: Planning, Priority Setting,Resource Allocation. McGraw-Hill.

Srdjevic, B., Medeiros, Y.D.P., Faria, A.S., 2004. An objective multi-criteria evaluationof water management scenarios. Water Resour. Manage 18, 35–54.

Storgard, J., Erdogan, I., Lappalainen, J., Tapaninen, U., 2012. Developing incidentand near missing reporting in the maritime industry-a case study on the BalticSea. Proc. Soc. Behav. Sci. 48, 1010–1021.

Tarelko, W., 2012. Origins of ship safety requirements formulated by InternationalMaritime Organization. Int. Symp. Saf. Sci. Technol. Proc. Eng. 45, 847–856.

Tzannatos, W., Kokotos, D., 2009. Analysis of accidents in Greek shipping during thepre-and post-ISM period. Mar. Policy 33, 679–684.

Vargas, L., 1982. Reciprocal matrices with random coefficients. Math. Model. 3, 69–81.

Yang, L.Z., Wang, J., Li, X.K., 2013. Maritime safety analysis in retrospect. MaritimePolicy Manag. 40 (3), 261–277.