Multi-Stage Multi-Objective Solid Transportation Problem for Disaster Response Operation with Type-2 Triangular Fuzzy Variables

Transcription

1 Multi-Stage Multi-Obective Solid Transportation Problem for Disaster Response Operation with Type- Triangular Fuzzy Variables Abhiit Baidya, Uttam Kumar Bera and Manoranan Maiti Abstract In this paper, for the first time we formulate and solve multi-stage solid transportation problem (MSSTP to minimize the total cost and time with type- fuzzy transportation parameters. During transportation period, loading, unloading cost and time, volume and weight for each item, limitation of volume and weight for each vehicle are normally imprecise and taken into account to formulate the models. To remove the uncertainty of the type- fuzzy transportation parameters from obective functions and constraints, we apply CV-Based reduction methods and generalized credibility measure. Disasters are unexpected situations that require significant logistical deployment to transport equipment and humanitarian goods in order to help and provide relief to victims and sometime this transportation is not possible directly from supply point to destination. Again, the availabilities at supply points and requirements at destinations are not own precisely due to disaster. For this reason, we formulate the multi-stage solid transportation problems under uncertainty (type- fuzzy. The models are illustrated with a numerical example. Finally, generalized reduced gradient technique (LINGO.3.0 software is used to solve the models. Keywords: Multi-stage solid transportation problem, type- fuzzy variable, CV-based reduction methods, generalized credibility, goal programming approach. 000 AMS Classification: 90B06, 90C05, 90C9. Introduction.. Literature review and the main work of the research: The transportation problem originally developed by Hitchcock [7] is one of the most common combinatorial problems involving constraints. The solid transportation problem (STP was first stated by Shell [8]. Haley [5, 6] developed the solution procedure of a solid transportation problem and made a comparison between the STP and the classical transportation problem. Geoffrion and Graves [9] were the first researchers studied on two-stage distribution problem. After that so many researchers(pirkul and Jayaraman [0], Heragu [7], Hindi et al. [], Syarif and Gen [], Amiri [3], Gen et al. [4] study two-stage TP. Mahapatra et al. [5] applied fuzzy multi-obective mathematical programming technique on a reliability optimization model. A type- fuzzy variable is a map from a fuzzy Department of Mathematics, National Institute of Technology, Agartala, Barala, Jirania,Tripura, India, Pin , abhiitnita@yahoo.in Corresponding Author. Department of Mathematics, National Institute of Technology, Agartala, Barala, Jirania,Tripura, India, Pin , bera uttam@yahoo.co.in Department of Applied Mathematics with oceanology and computer programming, Vidyasagar University, Midnapur-70, W. B.,India, mmaiti005@yahoo.co.in

2 possibility space to the real number space; it is an appropriate tool for describing type- fuzziness. The concept of a type- fuzzy set was first proposed in Qin et al. [] as an extension of an ordinary fuzzy set. Mitchell [0] used the concept of an embedded type- fuzzy number. Liang and Mendel [] proposed the concept of an interval type- fuzzy set. Karnik and Mendel [], Liu [3], Qin et al. [], Yang et al. [30], Liu et al. [9], Yang et al. [30] worked on type- fuzzy set. In the literature, data envelopment analysis (DEA technology was first proposed in []. Sengupta [] incorporated stochastic input and output variations; Banker [4] incorporated stochastic variables into DEA; Cooper et al. [5] and Land et al. [6] developed a chance-constrained programming to accommodate the stochastic variations in the data. Qin et al. [4] proposes three noble methods of reduction for a type- fuzzy variable. Here, we present in tabular form a scenario of literature development made on transportation problem in Table-. Table-: Some remarkable research works on TP/STP Author(s, Ref. Obective Nature Additional function Environments Solution Techniques Kundu et al. Multi-obective STP Multi-item fuzzy LINGO Yang and Feng Single-obective STP fixed charges stochastic Tabu search algorithm Kundu et al. Single-obective TP Fixed charge type- fuzzy variable LINGO Baidya et al. Single-obective STP Safety factor Fuzzy, Stochastic, Interval LINGO Gen et al. Single-obective TP Two-stage Deterministic Genetic algorithms Proposed Multi-obective STP Multi-stage Triangular type- fuzzy LINGO In spite of the above developments, there are so many gaps in the literature. Some of these omissions which are used to formulate the model with type- triangular fuzzy number are as follows: So many ([6], [7], [8], [3], [3], [33], solid transportation problems exist in the literature to minimize the total transportation cost only but nobody can formulate any STP to minimize the total transportation time, purchasing cost, loading and unloading cost at a time. In spite of the above developments, very few can minimized the time obective function which involves total transportation time, loading and unloading time at a time. Lots of two stage transportation problems ([9]-[5] exist in the literature where the transportation cost is minimized. But nobody formulate and solved a multi-stage multi-item multi-obective solid transportation problem to minimize the total cost which involves transportation cost, purchasing cost, loading and unloading cost and total time which involves transportation time, loading and unloading time. Sometimes, the value of the transportation parameters are not own to us precisely but at that time some imprecise data are own to us. For this reason, lots of researchers solved so many transportation problems with fuzzy (triangular fuzzy number, trapezoidal fuzzy number, type- fuzzy number, type- fuzzy number, interval type- fuzzy number etc. transportation parameters. But nobody solved any multi-stage multi-item multi-obective (cost and time STP with transportation parameters as type- triangular fuzzy number. So many STPs are developed in the literature to minimize the total transportation cost and time subected to the supply constraints, demand constraints and conveyances capacity constraints, budget constraint, safety constraint etc. but nobody formulated any STP subected to the weights constraints and volumes constraints during transportation.

3 3 In this paper, a multi-item multi-stage solid transportation problem is formulated and solved. Type- fuzzy theory is an appropriate field for research. To formulate the model, we consider unit transportation cost, time, supplies, demands, conveyances capacities, loading and unloading cost and time, volume and weights for each and every item, volume and weight capacities for each conveyances as type- triangular fuzzy variables. The obective functions for the respective transportation model is to minimize the total cost and time. To defuzzify the constraints and obective functions, we apply CV-based reduction method. The goal programming approach is used to solve the multi-obective programming problem. The deterministic problems so obtained are then solved by using the standard optimization solver - LINGO 3.0 software. We have provided numerical examples illustrating the proposed model and techniques. Some sensitivity analyzes for the model are also presented. The paper is organized as follows. Problem descriptions are includes in the section. In section 3 we briefly introduce some fundamental concepts. The assumptions and notations to construct the model are put in the section 4. In section 5, multi-stage multi-obective solid transportation model with type- triangular fuzzy variable is formulated. In section 6, we discuss about the methodology and defuzzification method that used to solve the model. A numerical example put in the section 7 is to illustrate the model numerically. The results of solving the model numerically are put in the section 8. A sensitivity analysis of the model is discussed in the section 9. In section 9, we discuss the results obtained by solving the numerical example. The comparisons of the work with the earlier research are discussed in the section 0. The conclusion and future extension of the research work are discussed in the section. The references which are used to prepare this manuscript are put in the last. In this work, we formulate and solve a multi-stage multi-item solid transportation problem to minimize the total cost and time under type- fuzzy environment. The real life applications of the research work are as follows: Basically to provide some relief to the survived peoples in disaster, we developed our MSSTP model. In our paper, we consider all the transportation parameters as type- fuzzy variables since after disaster, it is very difficult to define all transportation parameters precisely. Since due to disaster roads, bridges, towers etc. are damaged, thus it is not possible to survive the peoples smoothly with the help of direct transportation network. This is the reason to formulate an n-stage solid transportation model To get the permit of a vehicle is a difficult task for the owner of the vehicle. Some vehicles are permitted to driver on a particular state or country. This permit restricts us to carry the goods from one state to another or one country to another. So in the transportation period, it is important to load and unload the goods so many times at the destination centers (the destination centers are lies between supply points and customers.so to overcome these transportation difficulties we can apply this newly developed model... Motivations: The motivation for this research dated back to September 04, the Kashmir region witnessed disastrous floods across maority of its districts caused by torrential rainfall. The Indian administrated Jammu and Kashmir, as well as Azad Kashmir, Gilgit-Baltistan and Punab in Pakistan, were affected by these floods. By September 4, 04, nearly 77 people in India and 80 people in Pakistan died due to the floods and more than. million were affected by the floods. During this period, it is tedious to send the necessary foods to the survived peoples. For this reason, it is important to impose so many destination centers in between supply point and customers.

4 4. Problem Description: Disaster (earthquick, flood etc. is an extra ordinary situation for any country or state and to provide relief to the survived person which is a risky and tedious task to us. But at that time transportation is required to serve the foods, clothes etc. to the peoples. Also due to the disaster, it is not possible to deliver the necessary things directly to the survived people. It required some destination centers in between source point and survived peoples such that the total cost and time should be minimized. This also motivated us to formulate a multi-stage multi-obective solid transportation problem. The exact figure of the survived peoples due to disaster is not own to us exactly. For this reason, the transportation parameters are also remains unown to us. Since all the transportation parameters are not own to us precisely, so we consider the transportation parameters as type- triangular fuzzy variables. In this multi-stage transportation network, destination center for stage- is reduced to the supply point for stage- and destination center for stage- is reduced to the supply point for stage-3 and similarly the destination center for the stage-(n- is converted to the supply point to the stage-n. The pictorial representation of the multi-stage solid transportation problem is as follows: Figure. Multi-stage solid transportation network 3. Fundamental Concepts Definition. (Regular Fuzzy Variable Let Γ be the universe of discourse. An ample field [3] A on Γ is a class of subsets of Γ that is closed under arbitrary unions, intersections, and complements in Γ. Let P os : A [0, ] be a set function on the amplefield Γ. Pos is said to be a possibility measure [3] if it satisfies the following conditions: (p P os(ϕ 0 and P os(γ. (p For any subclass {A i i I} of A (finite, counter or uncountable, P os( i I (A i Sup i I P os(a i The triplet (Γ, A, P os is referred to as a possibility space, in which a credibility measure [33] is defined as Cr(A ( + P os(a P os(ac, A A. If (Γ, A, P os is a possibility space, then an m-ary regular fuzzy vector ξ (ξ, ξ,..., ξ m is defined as a membership map from Γ to the space [0, ] m in the sense that for every t (t, t,..., t m [0, ] m, one has {γ Γ ξ(γ t} {γ Γ ξ (γ t, ξ (γ t,..., ξ m (γ t m } A When m, ξ is called a regular fuzzy variable (RFV.

5 5 3.. Critical values for RFVs.. Definition.(Qin et al. [] Let ξ be an RFV. Then the optimistic CV of ξ, denoted by CV [ξ], defined as CV [ξ] Sup{α P os{ξ α}}, while the pessimistic CV of ξ, denoted by CV [ξ], is }{{} α [0,] defined as CV [ξ] Sup{α Nec{ξ α}}. }{{} α [0,] The CV of ξ, denoted by CV [ξ], is defined as CV [ξ] Sup{α Cr{ξ α}}. }{{} α [0,] Theorem. (Qin et al. [] Let ξ (r, r, r 3, r 4 be a trapezoidal RFV. Then we have (i The optimistic CV of ξ is CV [ξ] (ii The pessimistic CV of ξ is CV [ξ] (iii The CV of ξ is CV [ξ] r 4 +r 4 r 3. r +r r. r r +(r r, if r >, if r r 3 r 4 +(r 4 r 3, r Methods of reduction for type- fuzzy variables (CV-Based Reduction Methods. Due to the fuzzy membership function of a type- fuzzy number, the computation complexity is very high in practical applications. To avoid this difficulty, some defuzzification methods have been proposed in the literature (see [6-8]. In this section, we propose some new methods of reduction for a type- fuzzy variable. Compared with the existing methods, the new methods are very much easier to implement when we employ them to build a mathematical model with type- fuzzy coefficients. Let (Γ, A, P os be a fuzzy possibility space and ξ a type- fuzzy variable with a own secondary possibility distribution function µ ξ(x. To reduce the type- fuzziness, one approach is to give a representing value for RFV µ ξ(x. For this purpose, we suggest employing the CVs of P os{γ ξ(γ x} as the representing values. This methods the CV-based methods for the type- fuzzy variable ξ Theorem.(Qin et al. [] Let ξ be a type- triangular fuzzy variable defined as ξ ( r, r, r 3 ; θ l, θr. Then we have (i Using the optimistic CV reduction method, the reduction ξ of ξ has the following possibility distribution: (+θ r(x r (r, if x [r r +θ r(x r, r+r ] ( θ rx+θ rr r r+r (r µ ξ (x r +θ l (r x, if x [, r ] ( +θ rx θ rr +r 3 r 3 r +θ r(x r, if x [r, r+r3 ] (+θ r(r 3 x r, r+r3 3 r +θ r(r 3 x if x [, r 3 ] (ii Using the pessimistic CV reduction method, the reduction ξ of ξ has the following possibility distribution: x r r r +θ l (x r, if x [r, r+r ] x r r µ ξ (x r +θ l (r, r+r x if x [, r ] r 3 x r 3 r +θ l (x r, if x [r, r+r3 ] r 3 x r 3 r +θ l (r, r+r3 3 x if x [, r 3 ]

6 6 (iii Using the CV reduction method, the reduction ξ 3 of ξ has the following possibility distribution: (+θ r(x r r, if x [r r +θ r(x r, r+r ] ( θ l x+θ l r r r µ ξ3 (x r +θ l (r, r+r x if x [, r ] ( +θ l x θ l r +r 3 r 3 r +θ l (x r, if x [r, r+r3 ] (+θ r(r 3 x r, r+r3 3 r +θ r(r 3 x if x [, r 3 ] 3.3. Generalized credibility and its properties. Suppose ξ is a general fuzzy variable with the distribution µ. The generalized credibility measure Cr of the event {ξ α} is defined by Cr({ξ α} (Sup x Rµ(x + Sup x r µ(x Sup x<r µ(x, r R. Therefore, if ξ is normalized, it is easy to check that Cr(ξ α + Cr(ξ < α Sup x R µ ξ (x ; then Cr coincides with the usual credibility measure. The concept of independence for normalized fuzzy variables and its properties were discussed in [35]. In the following, we also need to extend independence to general fuzzy variables. The general fuzzy variables ξ, ξ, ξ..., ξ n are said to be mutually independent if and only if Cr{ξi B i, i,,...n} Min i ncr{ξi B i } for any subsets B i, i,,...n of R Like the α-optimistic value of the normalized fuzzy variable [36], the α-optimistic value of general fuzzy variables can be defined through the generalized credibility measure. Let ξ be a fuzzy variable (not necessary normalized. Then ξ Sup (α Sup{r Cr{ξ r} α}, α [0, ] is called the α-optimistic value of ξ, while ξ inf Inf{r Cr{ξ r} α}, α [0, ], is called the α pessimistic value of ξ. Theorem 3. (Qin et al. [] Let ξ i be the reduction of the type- fuzzy variable ξ i ( r, i r, i r 3; i θ l,i, θ r,i obtained by the CV reduction method for i,,...n. Suppose ξ, ξ,..., ξ n are mutually independent, and k i o for i,,...n. Case-I: If α (0, 0.5], then eqnarray Cr{ n i k iξ i t} α is equivalent to n i ( α + ( 4αθ r,i k i r i + αk i r i + ( 4αθ r,i Case-II: If α (0.5, 0.50], then eqnarray Cr{ n i k iξ i t} α is equivalent to n i ( αk i r i + (α + (4α θ l,i k i r i + ( 4αθ l,i Case-III: If α (0.50, 0.75], then eqnarray Cr{ n i k iξ i t} α is equivalent to n i (α k i r i 3 + (( α + (3 4αθ l,i k i r i + (3 4αθ l,i

7 7 Case-IV: If α (0.75, ], then eqnarray Cr{ n i k iξ i t} α is equivalent to n i (α + (4α 3θ r,i k i r i 3 + ( αk i r i + (4α 3θ r,i 3.4. Goal programming Method.. The goal programming method is used to solve the multi-obective programming problem (MOPP. A general MOPP is of the following form: (3. Find the values of L decision variables x, x,..., x L which minimizes F (x (f (x, f (x,..., f Q (x T subect to x X Where, X {x (x, x,..., x L such that g t (x 0, x l 0, t,,..., T ; l,,..., L} and f (x, f (x,..., f Q (x are Q( obective functions. The different steps of the goal programming method are as follows: Step-: Solve the multi-obective programming problem ( as a single obective problem using only one obective at a time ignoring the others, and determine the ideal obective vector, say f min, f min,..., fq min. Step-: Formulate the following GP problem using the ideal obective vector obtained is Step-, Min{ [(d + q p + (d q p ]} p q subect to f q (x + d + q d q fq min, d + q 0, d q 0, d + q d q 0(q,,..., Q, for all x X. Step-3: Now, solve the above single obective problem described in Step- by GRG method and obtain the compromise solution. 4. Notations and assumptions for the proposed model (i C ik, C q kk, C n q lmk nq Fuzzy unit transportation cost is to transport the q-th item from i-th plant to -th DC by k -th vehicle, -th plant to k-th DC k -th vehicle and l-th plant to m-th customer k n -th vehicle respectively. (ii t ik, q t kk, q t n lmk nq Fuzzy unit transportation time is to transport the q-th item from i-th plant to -th DC by k -th vehicle, -th plant to k-th DC k -th vehicle and l-th plant to m-th customer k n -th vehicle respectively. (iii x ik, q x kk, q xn lmk, nq xn ulk (n q, xn lmk nq Unown quantities which is to be transported from i-th plant to -th DC of q-th item by k -th vehicle for stage-, -th plant to k-th DC of q-th item by k -th vehicle for stage-, u-th plant to l-th DC of q-th item by k (n -th vehicle for stage-(n, and l-th plant to m-th customer of q-th item by k n -th vehicle for stage-n respectively.

8 8 (iv P C Fuzzy purchasing cost of q-th item at i-th source. (v (vi (vii (viii n i,,..., l Fuzzy loading cost at i-th plant of stage-, -th plant of stage- and l-th plant of stage-n respectively. UD, UD k,..., UD n m Fuzzy unloading cost at -th DC of stage-, k-th DC of stage- and m-th customer of stage-n respectively. O i, O,..., O n l Fuzzy loading time at i-th plant of stage-, -th plant of stage- and l-th plant of stage-n respectively. UT D, UT D k,..., UT D n m Fuzzy unloading time at -th DC of stage-, k-th DC of stage- and m-th customer of stage-n respectively. (ix y ik q {, if x ikq > 0 0, otherwise, y kk q {, if x kkq > 0 0, otherwise, y n lmk nq {, if x n lmq > 0 0, otherwise 5. Formulation of solid transportation problem with transportation parameters as type- triangular fuzzy variables Let us consider I supply points (or sources, J destination centers, K conveyances for stage- transportation; J supply points (or sources, K destination centers, K conveyances for stage- transportation; U supply points, L destination centers, k (n conveyances for stage-(n transportation; L supply points (or sources, M destination centers, K n conveyances for stage-n transportation. Also we consider that Q be the number of items which is to be transported from plants to DC by different modes of conveyances. (5. (5. I K Minf ( C ik q + i + UD + P C iqx ik q q i k K K + ( C kk q + + UD k + P C qx kk q +... q k k L M K n + ( C lmk n nq + n l + UD n m + P C n lqx n lmk nq q l m k n I K Minf ( t ik q + O i + UT D y ikq q i k K K + ( t kk q + O + UT D kykk q +... q k k L M K n + ( t n lmk nq + O n l + UT D n mylmk n nq q l m k n

9 9 (5.3 (5.4 (5.5 (5.6 (5.7 (5.8 (5.9 (5.0 (5. (5. (5.3 (5.4 (5.5 (5.6 (5.7 K k I K i k I q i I q i I q i K K k k K k x ik q ã iq, i,,..., I; q,,..., Q, x ik q b q,,,..., J; q,,..., Q, x ik q ẽ k, k,,..., K, w q x ik q W k, k,,..., K, q k q k q k M K n m k n L l k n K L ṽ q x ik q Ṽ k, k,,..., K, x kk q I K i k x ik q,,,..., J; q,,..., Q, x kk q b kqk,,...k; q,,..., Q, K x kk q ẽ k, k,,..., K, K w q x kk q W k, k,,..., K, K q l m L q l m L ṽ q x kk q Ṽ k, k,,..., K,... K U (n x n lmk nq u k (n x (n ulk (n q, l,,..., L; q,,..., Q, x n lmk nq b n mq, m,,...m; q,,..., Q, M x n lmk nq ẽ n k n, k n,,..., K n, M M q l m w q x n lmk nq W n k n, k n,,..., K n, ṽ q x n lmk nq Ṽ n k n, k n,,..., K n, x ik q 0, x kk q 0,..., x (n ulk (n q 0, xn lmk nq 0, for all i,, k, u, l, m, q, k, k, k (n, k n.

10 0 where, W k, W k, W k n n are the fuzzy weight capacity of k -th vehicle of stage-, k -th vehicle of stage-, k n -th vehicle of stage-n. Ṽk, Ṽ k, Ṽ k n n are the fuzzy volume capacity of k -th vehicle of stage-, k -th vehicle of stage-, k n -th vehicle of stage-n. w q, ṽ q are the fuzzy weight and volume of the q-th item. Also ã iq be the fuzzy availabilities of the q-th item at i-th source of stage-. b q, b kq, and b n mq are the fuzzy demands of q-th item at -th DC, k-th DC and m-th customer for stage-, stage- and stage-3 respectively. Also, ẽ k ẽ k, ẽ n k n are the fuzzy conveyances capacities of the k -th, k -th,k n -th conveyances for stage-, stage- and stage-n respectively. In this model formulation, we are to minimize two obective functions as total cost and time under supply, demand, conveyance capacity, weight and volume constraints. Here the first summation of the first obective indicates the total cost for stage- transportation. Similarly, second and last summation of the first obective function indicates the total cost for stage- and stage-n transportation respectively. Also the three summations of second obective denotes the total time in transportation respectively for stage-, stage- and stage-n respectively. We formulate the model in such a way that the goods are loaded at the supply point and it is unloaded at the DC for stage- transportation. Since due to disaster, it is not possible to move the vehicle directly to the survived people so after unloading at the first DC it again loaded to another vehicle and goes to the next DC and it is unloaded again in second DC for stage-. In this way the necessary goods are transported to the survived peoples or customers. For this reason, we impose the loading and unloading cost and time for each stage. Again purchasing cost is also imposed in our model. 6. Methodology and defuzzification technique used to solve the Model 6.. Methodology. The world has become more complex and almost every important real-world problem involves more than one obective. In such cases, decision makers find imperative to evaluate best possible approximate solution alternatives according to multiple criteria. To solve such multi-obective programming problem we apply goal programming method. Using CV -based reduction method and generalized credibility measure we find the deterministic form of type- fuzzy transportation parameters. Finally generalized reduced gradient technique (LINGO 3.0 optimization software is used to solve the developed model. 6.. Defuzzification. The deterministic form of the obective functions and constraints obtained by using CV -based reduction method and generalized credibility measure are as follows: (6. I K Cr{( ( C ik q + i + UD + P C iqx ik q q i k K K + ( C kk q + + UD k + P C qx kk q +... q k k L M K n + ( C lmk n nq + n l + UD n m + P C n lqx n lmk nq f } α c q l m k n

11 (6. I K Cr{( ( t ik q + O i + UT D yik q q i k K K + ( t kk q + O + UT D kykk q +... q k k L M K n + ( t n lmk nq + O n l + UT D n mylmk n nq f } α t q l m k n (6.3 (6.4 (6.5 (6.6 (6.7 (6.8 (6.9 (6.0 (6. (6. (6.3 Cr( Cr( K k I K i k Cr( Cr( I q i Cr( Cr( q i x ik q ã iq α avail., i,,..., I; q,,..., Q, x ik q b q α demand,,,..., J; q,,..., Q, I I q i K k q k x ik q ẽ k α con.cap., k,,..., K, w q x ik q W k α weight, k,,..., K, ṽ q x ik q Ṽ k α volume, k,,..., K, x kk q b kq α demand, k,,...k; q,,..., Q, K Cr( x kk q ẽ k α con.cap., k,,..., K, K Cr( w q x kk q W k α weight, k,,..., K, q k K Cr( ṽ q x kk q Ṽ k α volume, k,,..., K, L Cr( q k K l k n q l m... x n lmk nq b n mq α demand, m,,...m; q,,..., Q, L M Cr( x n lmk nq ẽ n k n α con.cap. v, k n,,..., K n,

12 (6.4 (6.5 L M Cr( w q x n lmk nq W k n n α weight, k n,,..., K n, q l m L M Cr( q l m ṽ q x n lmk nq Ṽ n k n α volume, k n,,..., K n, Let us consider α c, α t, α avail., α demand, α con.cap., α weight, α volume be the credibility level for cost, time, availabilities, demands, conveyances capacities, weights, volume respectively for stage-, stage-,...,stage-n. The crisp conversion of the constraints (6.-(6.5 are as follows: Minf + + Minf k I K i k K + + q i I q i I I q i I q i k q k k L K (S C + S ik q i + S UD + S P C iqx ik q K K (S C + S kk q + S UD + S k P C qx kk q +... M q l m k n I q i k q k k L K n (S Cn + S lmq K (S t + S ik q n l O i + S UD n m + S P C n lqx n lmk nq + S UT D y ikq K K (S t + S kk q O + S UT D kykk q +... M q l m k n K n (S t + S n lmq O n l x ik q Sã iq, i,,..., I; q,,..., Q, x ik q S b q,,,..., J; q,,..., Q, x ik q Sẽ k, k,,..., K, S wq x ik q S W, k,,..., K, k Sṽq x ik q SṼ k, k,,..., K, + S UT D n my n lmk nq

13 3 K k q k q k q k L K x kk q S b kq k,,...k; q,,..., Q, K x kk q Sẽ k, k,,..., K, K S wq x kk q S W, k,,..., K, k K l k n L q l m L q l m L q l m Sṽq x kk q SṼ k, k,,..., K,... x n lmk nq S bn mq, m,,...m; q,,..., Q, M x n lmk nq Sẽn, k n,,..., K n, M S wqx n S lmq W n, k n,,..., K n, M Sṽq x n lmk nq SṼ n, k n,,..., K n, Where S C, S ik q C S kk q, Cn, lmq S t, S t, ik q S t S kk q n lmq, S i, S n, S l UD, S UD S k, UD n m, SP C S iq, P C S q, P C n S lq, O, S i O, S O n, S l UT D, S UT D S k, UT D n Sã m,, iq,,,, Sẽ S b q S b kq S bn mq k, Sẽ k, Sẽn, S wq, S W, Sṽq, SṼ, S k k W, SṼ, S k k W n, SṼ n are equivalent crisp form of fuzzy parameters respectively and given as follows: C ik ( α+( 4α cθ r, r q C ik C +α cr q ik q (+( 4α cθ r, C ik q, if 0 < α c 0.5 S C ik q C ik ( α cr q +(α c+(4α c θ l, r C ik q (+( 4α cθ l, C ik q C ik q C ik (α c r q 3 +(( α c+(3 4α cθ l, r C ik q (+(3 4α cθ l, C ik q C ik (α c +(4α c 3θ r, r q C 3 +( α cr ik q (+(4α c 3θ r, C ik q, if 0.5 < α c 0.50 C ik q, if 0.50 < α c 0.75 C ik q, if 0.75 < α c

14 4 S i SUD S P C iq ( α c+( 4α cθ r, r i +α cr i i (+( 4α cθ r, ( α cr i +(α c+(4α c θ l, r i i (+( 4α cθ l, (α c r i 3 +(( α c+(3 4α cθ l, r i i (+(3 4α cθ l, (α c +(4α c 3θ r, r i 3 +( α cr i i (+(4α c 3θ r, i, if 0 < α c 0.5 UD ( α c+( 4α cθ r, UD r UD +α cr (+( 4α cθ r, UD UD ( α cr +(α c+(4α c θ l, UD r UD (+( 4α cθ l, UD UD (α c r3 +(( α c+(3 4α cθ l, UD r UD (+(3 4α cθ l, UD (α c +(4α c 3θ r, UD i, if 0.5 < α c 0.50 i, if 0.50 < α c 0.75 i, if 0.75 < α c UD r, if 0 < α c 0.5, if 0.5 < α c 0.50, if 0.50 < α c 0.75 UD 3 +( α cr (+(4α c 3θ r, UD P ( α c+( 4α cθ r, P C r C iq P iq +α C iq cr (+( 4α cθ r,, if 0.75 < α c P C iq, if 0 < α c 0.5 P ( α C iq P cr +(α c+(4α c θ l, P C r C iq iq (+( 4α cθ l, P (α C iq P c r3 +(( α c+(3 4α cθ l, P C r C iq iq (+(3 4α cθ l, P (α c +(4α c 3θ r, P C r C iq P iq 3 +( α C iq cr (+(4α c 3θ r, C kk ( α c+( 4α cθ r, r q C +α cr kk q (+( 4α cθ r, C kk q P C iq, if 0.5 < α c 0.50 P C iq, if 0.50 < α c 0.75 P C iq, if 0.75 < α c C kk q, if 0 < α c 0.5 S C kk q C kk ( α cr q +(α c+(4α c θ l, r C kk q (+( 4α cθ l, C kk q C kk q C kk (α c r q 3 +(( α c+(3 4α cθ l, r C kk q (+(3 4α cθ l, C kk q C kk (α c +(4α c 3θ r, r q C 3 +( α cr kk q (+(4α c 3θ r, C kk q, if 0.5 < α c 0.50 C kk q, if 0.50 < α c 0.75 C kk q, if 0.75 < α c

15 S S UD k ( α c+( 4α cθ r, r +α cr (+( 4α cθ r, ( α cr +(α c+(4α c θ l, r (+( 4α cθ l, (α c r3 +(( α c+(3 4α cθ l, r (+(3 4α cθ l, (α c +(4α c 3θ r, r 3 +( α cr (+(4α c 3θ r,, if 0 < α c 0.5, if 0.5 < α c 0.50, if 0.50 < α c 0.75, if 0.75 < α c UD ( α c+( 4α cθ r, UD r k UD +α cr k k (+( 4α cθ r, UD k UD ( α cr k +(α c+(4α c θ l, UD r UD k k (+( 4α cθ l, UD k UD (α c r k 3 +(( α c+(3 4α cθ l, UD r UD k k (+(3 4α cθ l, UD k UD r k (α c +(4α c 3θ r, UD k (+(4α c 3θ r,, if 0 < α c 0.5, if 0.5 < α c 0.50, if 0.50 < α c 0.75 UD 3 +( α cr k, if 0.75 < α c UD k C lmq n C lmq n ( α c+( 4α cθ r, Cn r +α cr lmq (+( 4α cθ r, Cn lmq, if 0 < α c S Cn lmq C lmq n ( α cr +(α c+(4α c θ l, Cn r lmq C n lmq (+( 4α cθ l, Cn lmq, if 0.5 < α c 0.50 C lmq n (α c r3 +(( α c+(3 4α cθ l, Cn r lmq C n lmq (+(3 4α cθ l, Cn lmq, if 0.50 < α c 0.75 S n l S UD n m C lmq n C lmq n (α c +(4α c 3θ r, Cn r3 +( α cr lmq (+(4α c 3θ r, Cn lmq, if 0.75 < α c ( α c+( 4α cθ r, n r n l l +α n cr l (+( 4α cθ r, ( α n cr l +(α c+(4α c θ l, n r n l l (+( 4α cθ l, (α n c r l 3 +(( α c+(3 4α cθ l, n r n l l (+(3 4α cθ l, (α c +(4α c 3θ r, n r n l l 3 +( α n cr l (+(4α c 3θ r, n l, if 0 < α c 0.5 n l, if 0.5 < α c 0.50 n l, if 0.50 < α c 0.75 n l, if 0.75 < α c ( α c+( 4α cθ r, UD n r UD n m UD m +α n cr m (+( 4α cθ r, UD ( α n cr m +(α c+(4α c θ l, UD n r UD n m m (+( 4α cθ l, UD (α n c r m 3 +(( α c+(3 4α cθ l, UD n r UD n m m (+(3 4α cθ l, (α c +(4α c 3θ r, UD n r UD n m UD m 3 +( α n cr m (+(4α c 3θ r, UD n m, if 0 < α c 0.5 UD n m, if 0.5 < α c 0.50 UD n m, if 0.50 < α c 0.75 UD n m, if 0.75 < α c

16 6 S t ik q S O i SUT D S t kk q t ik ( α t+( 4α tθ r, t r q t ik +α tr q ik q (+( 4α tθ r, t ik q t ik ( α tr q t ik +(α t+(4α t θ l, t r q ik q (+( 4α tθ l, t ik q t ik (α t r q t ik 3 +(( α t+(3 4α tθ l, t r q ik q (+(3 4α tθ l, t ik q t ik (α t +(4α t 3θ r, t r q 3 +( α tr ik q, if 0 < α t 0.5, if 0.5 < α t 0.50, if 0.50 < α t 0.75 t ik q (+(4α t 3θ r, t ik q ( α t+( 4α tθ r, O r O i +α O tr i i (+( 4α tθ r, O i ( α O tr i +(α t+(4α t θ l, O r O i i (+( 4α tθ l, O i (α O t r i 3 +(( α t+(3 4α tθ l, O r O i i (+(3 4α tθ l, O i (α t +(4α t 3θ r, O r O i 3 +( α O tr i i (+(4α t 3θ r, O i UT ( α t+( 4α tθ r, UT D r D UT +α D tr (+( 4α tθ r, UT D UT ( α D tr +(α t+(4α t θ l, UT D r UT D (+( 4α tθ l, UT D UT (α D t r3 +(( α t+(3 4α tθ l, UT D r UT D (+(3 4α tθ l, UT D UT (α t +(4α t 3θ r, UT D r D UT 3 +( α D tr (+(4α t 3θ r, UT D t kk ( α t+( 4α tθ r, t r q +α tr kk q (+( 4α tθ r, t kk q, if 0.75 < α t, if 0 < α t 0.5, if 0.5 < α t 0.50, if 0.50 < α t 0.75, if 0.75 < α t, if 0 < α t 0.5, if 0.5 < α t 0.50, if 0.50 < α t 0.75, if 0.75 < α t t kk q t kk ( α tr q t kk +(α t+(4α t θ l, t r q kk q (+( 4α tθ l, t kk q t kk (α t r q t kk 3 +(( α t+(3 4α tθ l, t r q kk q (+(3 4α tθ l, t kk q t kk (α t +(4α t 3θ r, t r q t kk 3 +( α tr q kk q (+(4α t 3θ r, t kk q, if 0 < α t 0.5, if 0.5 < α t 0.50, if 0.50 < α t 0.75, if 0.75 < α t

17 S O S UT D k ( α t+( 4α tθ r, O r O +α O tr (+( 4α tθ r, O ( α O tr +(α t+(4α t θ l, O r O (+( 4α tθ l, O (α O t r3 +(( α t+(3 4α tθ l, O r O (+(3 4α tθ l, O, if 0 < α t 0.5 (α t +(4α t 3θ r, O r O 3 +( α O tr (+(4α t 3θ r, O UT ( α t+( 4α tθ r, UT D r D k UT +α D tr k k (+( 4α tθ r, UT D k UT ( α D tr k +(α t+(4α t θ l, UT D r UT D k k (+( 4α tθ l, UT D k UT (α D t r k 3 +(( α t+(3 4α tθ l, UT D r UT D k k (+(3 4α tθ l, UT D k UT (α t +(4α t 3θ r, UT D r D k UT 3 +( α D tr k k (+(4α t 3θ r, UT D k l n lmq l n lmq ( α t+( 4α tθ r, ln r +α tr lmq, if 0.5 < α t 0.50, if 0.50 < α t 0.75, if 0.75 < α t, if 0 < α t 0.5, if 0.5 < α t 0.50, if 0.50 < α t 0.75, if 0.75 < α t (+( 4α tθ r, ln lmq, if 0 < α t S ln lmq l n lmq l n lmq ( α tr +(α t+(4α t θ l, ln r lmq (+( 4α tθ l, ln lmq, if 0.5 < α t 0.50 l n lmq l n lmq (α t r3 +(( α t+(3 4α tθ l, ln r lmq (+(3 4α tθ l, ln lmq, if 0.50 < α t 0.75 S O n l S UT D n m l n lmq (α t +(4α t 3θ r, ln r3 +( α tr lmq l n lmq (+(4α t 3θ r, ln lmq, if 0.75 < α t ( α t+( 4α tθ r, O n r O n l l +α O n tr l (+( 4α tθ r, ( α O n tr l +(α t+(4α t θ l, O n r O n l l (+( 4α tθ l, (α O n t r l 3 +(( α t+(3 4α tθ l, O n r O n l l (+(3 4α tθ l, (α t +(4α t 3θ r, O n r O n l l 3 +( α O n tr l (+(4α t 3θ r, ( α t+( 4α tθ r, D n r D n m m +α D n tr m (+( 4α tθ r, ( α D n tr m +(α t+(4α t θ l, D n r D n m m (+( 4α tθ l, (α D n t r m 3 +(( α t+(3 4α tθ l, D n r D n m m (+(3 4α tθ l, (α t +(4α t 3θ r, D n r D n m m 3 +( α D n tr m (+(4α t 3θ r, O n l, if 0 < α t 0.5 O n l, if 0.5 < α t 0.50 O n l, if 0.50 < α t 0.75 O n l, if 0.75 < α t D n m, if 0 < α t 0.5 D n m, if 0.5 < α t 0.50 D n m, if 0.50 < α t 0.75 D n m, if 0.75 < α t

18 8 Sã iq S b q rã iq rã iq ( α avail. +( 4α avail. θ r,ã +α avail. iq (+( 4α avail. θ r,ã, if 0 < α avail. 0.5 iq rã iq rã iq ( α avail. +(α avail. +(4α avail. θ l,ã iq (+( 4α avail. θ l,ã, if 0.5 < α avail iq rã iq rã iq (α avail. 3 +(( α avail. +(3 4α avail. θ l,ã iq (+(3 4α avail. θ l,ã, if 0.50 < α avail iq rã iq rã iq (α avail. +(4α avail. 3θ r,ã 3 +( α avail. iq (+(4α avail. 3θ r,ã, if 0.75 < α avail. iq ( α demand +( 4α demand θ r, b q r b q +α demand r b q (+( 4α demand θ r, b q, if 0 < α demand 0.5 ( α demand r b q +(α demand +(4α demand θ l, b q r b q (+( 4α demand θ l, b q, if 0.5 < α demand 0.50 (α demand r b q 3 +(( α demand +(3 4α demand θ r b q l, b q (+(3 4α demand θ l, b, if 0.50 < α demand 0.75 q Sẽ k S wq (α demand +(4α demand 3θ r b q r, b 3 +( α demand r b q q (+(4α demand 3θ r, b, if 0.75 < α demand q ( α con.cap.+( 4α con.cap.θ r,tildee r tildee k +α tildee k con.cap.r k (+( 4α con.cap.θ r,tildee k, if 0 < α con.cap. 0.5 ( α tildee k con.cap.r +(α con.cap.+(4α con.cap. θ l,tildee r tildee k k (+( 4α con.cap.θ l,tildee k, if 0.5 < α con.cap (α tildee k con.cap. r 3 +(( α con.cap.+(3 4α con.cap.θ l,tildee r tildee k k (+(3 4α con.cap.θ l,tildee k, if 0.50 < α con.cap (α con.cap. +(4α con.cap. 3θ r,tildee r tildee k 3 +( α tildee k con.cap.r k (+(4α con.cap. 3θ r,tildee k, if 0.75 < α con.cap. ( α weight +( 4α weight θ r, wq r wq +α weightr wq (+( 4α weight θ r, wq, if 0 < α weight 0.5 ( α weight r wq +(α weight+(4α weight θ l, wq r wq (+( 4α weight θ l, wq, if 0.5 < α weight 0.50 (α weight r wq 3 +(( α weight+(3 4α weight θ l, wq r wq (+(3 4α weight θ l, wq, if 0.50 < α weight 0.75 (α weight +(4α weight 3θ r, wq r wq 3 +( α weightr wq (+(4α weight 3θ r, wq, if 0.75 < α weight

19 S W k W k W ( α weight +( 4α weight θ r, W r k +α weight r k (+( 4α weight θ r, W k, if 0 < α weight 0.5 W k ( α weight r +(α weight +(4α weight θ l, W r k W k (+( 4α weight θ l, W k, if 0.5 < α weight 0.50 W k (α weight r 3 +(( α weight +(3 4α weight θ l, W r k (+(3 4α weight θ l, W k, if 0.50 < α weight 0.75 W k 9 Sṽq W k (α weight +(4α weight 3θ r, W r 3 +( α weight r k ( α volume +( 4α volume θ r,ṽq rṽq (+(4α weight 3θ r, W k, if 0.75 < α weight +α volumerṽq (+( 4α volume θ r,ṽq, if 0 < α volume 0.5 ( α volume rṽq +(α volume+(4α volume θ l,ṽq rṽq (+( 4α volume θ l,ṽq, if 0.5 < α volume 0.50 (α volume rṽq 3 +(( α volume+(3 4α volume θ l,ṽq rṽq (+(3 4α volume θ l,ṽq, if 0.50 < α volume 0.75 (α volume +(4α volume 3θ r,ṽq rṽq 3 +( α volumerṽq (+(4α volume 3θ r,ṽq, if 0.75 < α volume ( α volume +( 4α volume θ r, Ṽ k r Ṽ k Ṽ k +α volume r (+( 4α volume θ r, Ṽ k, if 0 < α volume 0.5 W k SṼ k S b kq ( α volume r Ṽ k Ṽ +(α volume +(4α volume θ l, Ṽ k k r (+( 4α volume θ l, Ṽ k, if 0.5 < α volume 0.50 (α volume r Ṽ k Ṽ 3 +(( α volume +(3 4α volume θ l, Ṽ k k r (+(3 4α volume θ l, Ṽ k, if 0.50 < α volume 0.75 (α volume +(4α volume 3θ r, Ṽ k r Ṽ k Ṽ k 3 +( α volume r (+(4α volume 3θ r, Ṽ k, if 0.75 < α volume b ( α demand +( 4α demand θ r, b r kq +α demand r kq b kq (+( 4α demand θ r, b kq, if 0 < α demand 0.5 b ( α demand r kq +(α demand +(4α demand θ l, b r kq b kq (+( 4α demand θ l, b kq, if 0.5 < α demand 0.50 b (α demand r kq 3 +(( α demand +(3 4α demand θ l, b r kq b kq (+(3 4α demand θ l, b kq, if 0.50 < α demand 0.75 b (α demand +(4α demand 3θ r, b r kq 3 +( α demand r kq b kq (+(4α demand 3θ r, b kq, if 0.75 < α demand

20 0 Sẽ k rẽ k ( α con.cap.+( 4α con.cap.θ k r,ẽ +α con.cap.rẽ k (+( 4α con.cap.θ r,ẽ k, if 0 < α con.cap. 0.5 con.cap.rẽ k ( α k +(α con.cap.+(4α con.cap. θ l,ẽ rẽ k (+( 4α con.cap.θ l,ẽ k, if 0.5 < α con.cap con.cap. rẽ k (α k 3 +(( α con.cap.+(3 4α con.cap.θ l,ẽ rẽ k (+(3 4α con.cap.θ l,ẽ k ẽ k, if 0.50 < α con.cap rẽ k (α con.cap. +(4α con.cap. 3θ k r,ẽ 3 +( α con.cap.rẽ k (+(4α con.cap. 3θ r,ẽ k, if 0.75 < α con.cap. W k ( α weight +( 4α weight θ r, W r +α weight r k (+( 4α weight θ r, W k, if 0 < α avail. 0.5 W k S W k W k ( α weight r +(α weight +(4α weight θ l, W r k W k (+( 4α weight θ l, W k, if 0.5 < α avail W k (α weight r 3 +(( α weight +(3 4α avail. θ l, W r k (+(3 4α weight θ l, W k, if 0.50 < α avail W k W k (α weight +(4α weight 3θ r, W r 3 +( α weight r k (+(4α weight 3θ r, W k, if 0.75 < α avail. ( α volume +( 4α volume θ r, Ṽ k r Ṽ k Ṽ k +α volume r (+( 4α volume θ r, Ṽ k, if 0 < α volume 0.5 W k SṼ k S bn mq ( α volume r Ṽ k Ṽ +(α volume +(4α volume θ l, Ṽ k k r (+( 4α volume θ l, Ṽ k, if 0.5 < α volume 0.50 (α volume r Ṽ k Ṽ 3 +(( α volume +(3 4α volume θ l, Ṽ k k r (+(3 4α volume θ l, Ṽ k, if 0.50 < α volume 0.75 (α volume +(4α volume 3θ r, Ṽ k r Ṽ k Ṽ k 3 +( α volume r (+(4α volume 3θ r, Ṽ k, if 0.75 < α volume ( α demand +( 4α demand θ r, bn r b n mq mq +α demand r b n mq (+( 4α demand θ r, bn, if 0 < α demand 0.5 mq ( α demand r b n mq +(α demand +(4α demand θ l, bn mq r b n mq (+( 4α demand θ l, bn mq, if 0.5 < α demand 0.50 (α demand r b n mq 3 +(( α demand +(3 4α demand θ l, bn r b n mq mq (+(3 4α demand θ l, bn, if 0.50 < α demand 0.75 mq (α demand +(4α demand 3θ r, bn mq r b n mq 3 +( α demand r b n mq (+(4α demand 3θ r, bn mq, if 0.75 < α demand

21 Sẽn rẽn n ( α con.cap.+( 4α con.cap.θ r,ẽ n +α con.cap.rẽ (+( 4α con.cap.θ r,ẽ n, if 0 < α con.cap. 0.5 con.cap.rẽn n ( α +(α con.cap.+(4α con.cap. θ l,ẽ n rẽ (+( 4α con.cap.θ l,ẽ n, if 0.5 < α con.cap con.cap. rẽn n (α 3 +(( α con.cap.+(3 4α con.cap.θ l,ẽ n rẽ (+(3 4α con.cap.θ l,ẽ n, if 0.50 < α con.cap rẽn n (α con.cap. +(4α con.cap. 3θ r,ẽ n 3 +( α con.cap.rẽ (+(4α con.cap. 3θ r,ẽ n, if 0.75 < α con.cap. W ( α weight +( 4α weight θ r, W n r n +α weight r W n (+( 4α weight θ r, W n, if 0 < α weight 0.5 S W n W n ( α weight r +(α weight +(4α weight θ l, W n r W n (+( 4α weight θ l, W n, if 0.5 < α weight 0.50 W n (α weight r 3 +(( α weight +(3 4α weight θ l, W n r (+(3 4α weight θ l, W n, if 0.50 < α weight 0.75 W n W (α weight +(4α weight 3θ r, W n r n 3 +( α weight r (+(4α weight 3θ r, W n, if 0.75 < α weight ( α volume +( 4α volume θ r, Ṽ n r Ṽ n Ṽ n +α volume r (+( 4α volume θ r, Ṽ n W n, if 0 < α volume 0.5 SṼ n ( α volume r Ṽ n Ṽ n +(α volume +(4α volume θ l, Ṽ n r (+( 4α volume θ l, Ṽ n (α volume r Ṽ n Ṽ n 3 +(( α volume +(3 4α volume θ l, Ṽ n r (+(3 4α volume θ l, Ṽ n (α volume +(4α volume 3θ r, Ṽ n r Ṽ n Ṽ n 3 +( α volume r (+(4α volume 3θ r, Ṽ n, if 0.5 < α volume 0.50, if 0.50 < α volume 0.75, if 0.75 < α volume 7. Numerical Example A firm produces two types of food as Bread and Biscuit and stored at two plants which are the supply points of our problem. The goods are delivered to two destination centers (DCs from these supply points then finally these products are transported to the final destination centers or customers or survived peoples on disaster via the first DCs. That is the transportation happened in two stages. Due to disaster, the requirements, availabilities and other transportation parameters are not own to us precisely. For this reason, we consider all the transportation parameters as type- triangular fuzzy numbers. The type- triangular fuzzy inputs for unit transportation costs and times, availabilities, demands, conveyances capacities, purchasing cost, loading and unloading cost and time, weights and volumes etc. for stage- and stage- are as follows: Type- fuzzy unit transportation cost, time for stage- and stage-: C (,, 4;.4,.6, C (, 3, 4;.,.3, C (, 3, 4;.,.3, C (, 4, 6;.,., C (3, 5, 6;.6,.7, C (4, 5, 6;.3,.5, C (3, 5, 6;.3,., C (4, 6, 7;.,.5, C (3, 4, 7;.4,.6,C (3, 4, 6;.,.8,C (, 5, 8;.,.7,C (, 4, 6;.5,., C (, 5, 7;.6,.9,C (, 3, 9;.3,.5,C (3, 7, 9;.3,.9,C (5, 6, 8;.4,.5,

22 t (, 3, 5;.4,.6,t (3, 4, 7;.7,.9,t (7, 9, ;.9,,t (, 4, 6;.,., t (7, 0, 3;.6,.9,t (5, 7, 8;.8,,t (4, 5, 8;.8,.3,t (6, 7, 9;.7,.5, t (5, 9, 3;.,.8,t (5, 7, 9;.8,.4,t (5, 8, 9;.9,.9,t (4, 5, 6;.5,.7, t (4, 6, 9;.4,.7,t (, 3, 9;.3,.5,t (3, 7, 0;.3,.9,t (5, 6, 8;.4,.5, C (8, 9, ;.4,.6,C (, 3, 4;.,,C (, 3, 4;.,.3,C (3, 5, 6;.,.7, C (3, 5, 6;.6,.9,C (4, 5, 6;.3,.5,C (3, 5, 6;.3,.,C (4, 6, 7;.6,.5, C (3, 4, 7;.4,.6,C (4, 5, 7;.,, C (, 5, 8;.,.7,C (3, 4, 7;.7,.9, C (, 5, 7;.6,.9,C (, 7, 9;.3,.5,C (6, 7, 9;.3,.9,C (5, 7, 8;.8,, t (, 4, 8;.6,.7,t (3, 5, 7;.4,.8,t (, 3, 4;.6,.7,t (, 4, 6;.,., t (7, 8, ;.,.,t (3, 7, 9;.8,.,t (, 5, 9;.8,.9,t (, 6, 8;.7,.5, t (3, 9, ;.3,.8,t (, 3, 4;.5,.,t (3, 4, ;.7,.4,t (4, 5, 6;.7,.3, t (3, 6, 8;.3,.4, t (, 6, 9;.,.8, t (7, 9, 3;.,.7, t (, 6, 9;.9,.. Table-: Type- fuzzy availabilities, demands, conveyances capacities, loading and unloading time and cost, weights and volume for stage- and stage- Availabilities ã (60, 66, 67;.,.5,ã (54, 56, 60;.,.,ã (4, 47, 55;.,.4,ã (47, 53, 55;.5,.6 Demands for stage- b (9, 6, 30;.,.3, b (, 4, 5;.,.3, b (0,, ;.7,., b (, 3, 5;.5,. Demands for stage- b (8, 0, 3;.,.3, b (7, 8, 5;.,.3, b (5, 6, 7;.,.4, b (, 4, 6;.9,.3 Conveyances Capacities ẽ (5, 54, 56;.,.3,ẽ (53, 55, 57;.6,.9,ẽ (4, 44, 49;.8,.3,ẽ (45, 49, 50;.4,.9 Loading Cost (, 4, 6;.,.3, (5, 6, 7;.,.3, (5, 6, 9;.,.6, (, 9, 0;.3,.4 Unloading Cost UD (, 3, 4;.4,.6, UD (3, 7, 8,.6;.7,, UD (, 3, 5;.4,.9, UD (6, 8, 9;.6, Loading Time O (3, 7, 9;.,.3, O (,, 3;.,., O (4, 6, 8;.,.3, O (, 3, 6;.,. Unloading Time UT D (3, 7, 0;.3,.8, UT D (, 9, ;.,.4, UT D (, 4, 5;.,.5, UT D (4, 5, 6;.5,.6 Purchasing Cost P C (0,, ;.,.4, P C (3, 5, 6;.,.4, P C (,, 3;.,., P C (4, 6, 7;.9,.3 Weights capacity W (85, 90, 5;.6,.7, W (88, 30, 400;.7,.8, W (30, 30, 33;.,.9, W (368, 340, 345;.6,.7 Volumes capacity Ṽ (334, 360, 370;.,.3,Ṽ (3, 94, 370;.6,.7, Ṽ (393, 395, 399;.6,.8,Ṽ (353, 354, 356;.8,.9 weight,volume of items w (, 5, 7;.4,.5, w (, 4, 5;.7,.9,ṽ (,, 6;.,.3,ṽ (, 3, 8;.,.5 8. Results The fuzzy multi-stage STP is converted to its equivalent crisp problem by using CV-based reduction method and generalized credibility measure. Then using LINGO.3.0 optimization software, we obtain the optimal solution of the deterministic STP. The values of the credibility level for the transportation parameters are sometime lies in the interval (0, 0.5] or (0.5, 0.5] or (0.5, 0.75] or (0.75, ]. For this reason, we obtain the optimal solution of the newly developed model with the four limitations of the credibility level. A sensitivity analysis is taken into consideration to show the change of the optimal values of the obective functions and the transported amounts with respect to the credibility level of availabilities, demands and conveyances capacities.

23 3 Table-3: Changes of optimum cost and transported amount for different credibility levels Credibility Level Item- Item- Item- Item- Stage- Stage- Opt. cost Opt. time α c 0.07 x.0 x 8.58 α t 0.0 x.90 x 7.9 α avail. 0.3 x 0.0 x 5.8 α demand x.0 x 6.88 α con.cap. 0.9 x 5.49 α weight 0. x 0.06 α volume α c 0.6 x 3.8, x, α t 0.9 x 6.9, x 3.5, α avail. 0.3 x.45, x 5.59, α demand x 4.30, x 4.7, α con.cap x 3.68, x.04, α weight 0.37 x 7.0, x 4.7 α volume α c 0.56 x 4.58, x 5.87, α t 0.59 x 5.87, x 4.84, α avail. 0.6 x 4.56, x 9.74, α demand x 6.64, x 4.43, α con.cap x 3.4 x 9.48, α weight 0.67 x 4.36, α volume 0.69 x.33. α c 0.76 x 4.04, x.6 α t 0.79 x 9.0, x 6.75, α avail x 0.7, x.57, α demand x.85, x 3.95, α con.cap x 7.6, x 6.4, α weight 0.95 x 7.0 x 6.76, α volume 0.98 x 9.4

24 4 8.. Particular Case. Let us consider, the credibility level for costs, times, availabilities, demands, conveyances capacities, weights and volume are all equal. i.e., α c α t α avail. α demand α con.cap. α weight α volume α, say. Table-4: Optimal results of the model with same credibility level Credibility Level Item- Item- Item- Item- Stage- Stage- Opt. cost Opt. time x.3, x 8.95 x.4, x 3.37, α x 0.44, x.46, x 5.09, x 9.0, x 5.47, x.9 x 5.70, x.6, x 4.90, x 3.9, α x 9.88, x 0.9, x 5.37, x 4.4, x 30.79, x 8.3, x x 5.84, x 6.49, x 6.49, x 4.8, α x 5.9, x 5.5, x 4.93, x 6., x 3.83, x 6.6, x 9.8, x x 8.0, x.87, x 8.49, x 3.67, α x 6.6, x.7, x 8.69, x 3.4, x 3., x 6.63, x.3, x Sensitivity Analysis of the availabilities and demands of the model. We ow that the sensitivity analysis is used to analyze the outputs with the given inputs data. For this reason, in Table-5 and -6 we analyze some inputs data and outputs as sensitivity analysis. Basically the minimization of cost obective and time obective in the STP depend on the values of the transportation parameters such as unit transportation costs, times, demands, supplies etc.

25 5 Table-5: Sensitivity analysis on availabilities α c α t α avail. α demand α con.cap. α weight α volume Opt. cost Opt. time Item- Item Table-6: Sensitivity analysis on demands α c α t α avail. α demand α con.cap. α weight α volume Opt. cost Opt. time Item- Item

26 Pictorial representation of the sensitivity analysis. The Pictorial representation of the sensitivity analysis are shown in the figure- - figure-7 and those are given below: Figure. Change of total optimum cost and time with Credibility level of availability, α avail. (0, 0.5] Figure 3. Change of total optimum cost and time with Credibility level of availability α avail. (0.5, 0.50]

27 7 Figure 4. Change of total optimum cost and time with Credibility level of availability, α avail. (0.50, 0.75] Figure 5. Change of total optimum cost and time with Credibility level of availability, α avail. (0.t5, ]

28 8 Figure 6. Change of total optimum cost and time with Credibility level of demand, α demand (0, 0.5] Figure 7. Change of total optimum cost and time with Credibility level of demand, α demand (0.5, 0.50] 9. Discussion Since the credibility level of the availabilities, demands, conveyances capacities, weights and volumes for each transported item and each vehicle are different, so after taking the variation of each transportation parameters, we obtained lots of results of our STP model where all the transportation parameters are type- fuzzy variables and which are discussed below: Following Table-3, we see that the least amount of total cost and time are and units respectively and these are obtained when the credibility level of the transportation parameters lies within the interval (0, 0.5]. Again, in Table-4, we put some optimal results which are obtained by taking the credibility level of all the transportation parameters are equal. After careful investigation, we found that the total cost and time are least when the credibility levels are same and it is lies in the interval (0, 0.5]. In Table-5 we have the following: (i When credibility level of the availabilities increases with the limit (0, 0.5], then the values of the cost obective also increases and the time obective are sometime increases and decreases. The

29 9 Figure 8. Change of total optimum cost and time with Credibility level of demand, α demand (0.50, 0.75] Figure 9. Change of total optimum cost and time with Credibility level of demand, α demand (0.75, ] increase or decrease of the total time within the variation of the credibility level is also significant i.e., if we change the credibility level then the allocations are changed and for this reason, it is happening. (ii From the third row of the Table-5, we obtained some optimal results of the obectives where the credibility level of availabilities are lies within (0.5, 0.5]. Due to change of credibility level of availabilities, sometimes the value of total cost are increased and sometimes decreased but there is no significant change in time obective. This is found when credibility level increases within the range (0.5, 0.5]. (iii The value of the cost obective increases when we increase the credibility level of availabilities within the range (0.5, 0.75] but there are some random changes found in the time obective function. (iv When credibility level of the availabilities increases within the limit (0.75, ], then the value of the obectives and transported amounts (item- and item- are increased. Again if we can change the value of the credibility level of the demand, then we found some significant changes on the obective functions as well as transported amounts. From Table-6, it is seen