CN103164747A - Battle field first-aid repair resource reorganization optimized decision method - Google Patents

Abstract

The invention relates to a battle field first-aid repair resource reorganization optimized decision method which is used for resource reorganization decisions of multi-task and multi-resource requirements and distributed type multi-resource-point supply. Firstly, a common battle field first-aid repair resource reorganization decision is initialized through a constraint propagation method, and inconsistent first-aid repair task variables and reorganization resource variables are reduced. Then according to requirement conflicts of first-aid repair tasks and first-aid repair resources in first-aid repair time, resource reorganization decisions are divided into three categories of non-conflict, complete-conflict and incomplete-conflict, on the basis that first-aid repair resource points are the minimum, flexibility of remained first-aid repair resources is the largest, the total consumption rate is the minimum and other elicitation rules, mixed algorithms which are minimum in total consumption rate and based on a binary discrete particle swarm algorithm are respectively designed according to category characteristics of different decisions. Finally, reorganization time value assignment strategies which begin at the same time, are closed at the same time, and are finished on time and the like are designed. The battle field first-aid repair resource reorganization optimized decision method aims at solving multi-task resource requirement conflicts in combined guarantee of existing systematic confrontation, so that a set of optimization value assignment scheme of first-aid repair tasks and reorganization resources and reorganization time can be generated in systematic and complete modes, and theoretical support and decision basis are provided for decision staff.

Description

A kind of BDAR both resource restructuring Study on Decision-making Method for Optimization

Technical field

the present invention relates to a kind of BDAR both resource restructuring decision-making technique, after the target of establishment system repairing maximizing the benefits, according to finding out an optimum solution that satisfies constraint, constraint condition comprises the time-constrain of continuous domain and the resource constraint of discrete domain, it is a kind of Combinatorial Optimization optimization problem with background meaning, in conjunction with definite type heuritic approaches such as total resources consumption rate minimum and resource flexibility minimums, and population equiprobability type heuritic approach, can be to the repairing task, recombine resource, the variable assignments such as reorganization time are optimized decision-making, has application prospect preferably in joint protection in the reply confrontation between systems.

Background technology

Combinatorial optimization is to go to seek optimum layout, grouping, order or the screening etc. of discrete event by the research of mathematical method, be a branch that classics are important of operational research, the problem of studying relates to infotech, management, transportation, production scheduling network and military field.In general the target of combinatorial optimization problem is to concentrate from the feasibility solution of combinatorial problem to obtain optimum solution.For the various combination optimization problem, do not have general high precision and the decision-making technique that has concurrently at a high speed and optimized algorithm, for the heuritic approach of different problem independent design, or be the feasibility thinking of finding the solution combinatorial optimization problem in conjunction with the hybrid algorithm of many algorithms.BDAR both resource restructuring decision-making has very strong Military Application background, requires the decision-maker within the limited time, obtains as far as possible more excellent resource restructuring scheme, satisfies the repairing mission requirements.In general the determinacy heuritic approach has computing velocity preferably, but the increase with variable assignment number of combinations, the precision of calculating is relatively poor, and probabilistic heuritic approach is when finding the solution the Constrained combinatorial optimization problem, due to issuable a large amount of infeasible solutions, can have influence on computational accuracy and computing velocity.The BDAR both resource restructuring decision-making technique that the design relates to, at first according to repairing resource and the constraint of repairing time, variable is carried out initialization by consistency detecting method, approximately subtract inconsistent variable, effectively reduce the dimension of general repairing resource restructuring decision problem, reduce the number of combinations of variable assignments.On initialized basis, according to different repairing tasks in the reorganization time constraint, the conditions of demand of repairing resource are classified, and according to the different corresponding optimized algorithms of classification characteristics design, the characteristics of the high computational accuracy of the high computing velocity of definite type heuritic approach such as minimum and resource flexibility minimum and particle cluster algorithm in conjunction with the total resources consumption rate, can obtain at short notice more excellent result of calculation, very strong theory significance and using value are arranged when determining the BDAR both decision scheme.

Summary of the invention

The present invention be directed to modern confrontation between systems joint protection characteristics and demand, the assignment decision-making technique of the how to confirm repairing task that proposes, recombine resource, reorganization time, the method is applicable to multitask, many resources, the resource restructuring of constrained concurrent repairing mission requirements.

BDAR both resource restructuring decision-making technique of the present invention comprises the following steps:

(1) adopt the modeling method of constraint satisfaction problem to carry out modeling to BDAR both resource restructuring decision-making, method by constraint propagation, according to repairing resource constraint and repairing time-constrain, the consistance of rushing to repair task is detected respectively, according to the constraint of repairing time, recombine resource is carried out consistency detection.

(2) according to the quantity constraint of repairing resource kind and the constraint of repairing time, by the method for constraint propagation, problem is carried out initialization, check repairing task is about repairing resource and repairing time consistency.The repairing mission requirements of recombinating in the reorganization time that after initialization, each repairing task has at least one group of recombine resource variable assignments combination to limit point.Simultaneously each repairing task supply centre provide certain repairing resource of a measurement unit can at least a repairing task.

(3) require in the repairing time according to different task, between the repairing task, the situation of the conflict of demand repairing resource will rush to repair the resource restructuring decision-making and will be divided into following three classes: rush to repair the resource restructuring decision-making without conflict, conflict repairing resource restructuring decision-making and conflict repairing resource restructuring decision-making fully fully, and model is set up in classification.

(4) according to rush to repair resource restructuring decision model characteristics without conflict, the repairing task has at least a resource provisioning point that all kinds of resources of repairing required by task can be provided within the restriction repairing time, belongs to " drug on the market " or " balance between supply and demand ".Giving the assignment of repairing assignment decisions variable is 1, and the task of namely rushing to repair all responds.Establish on this basis two-phase problem " the repairing resource points is minimum " and " residue repairing resource flexibility is maximum ", and determine respectively the assignment of recombine resource variable by the heuristic search algorithm of correspondence.

(5) according to the repairing resource restructuring decision model characteristics of conflicting fully, designed the mixture model method for solving based on resource wastage in bulk or weight rate and particle cluster algorithm, determine the assignment of repairing task variable, again according to the heuristic of residue repairing resource flexibility maximum, determine the assignment of recombine resource variable on this basis.

(6) rush to repair the characteristics of resource restructuring decision model according to not exclusively conflicting, repairing resource contention and flexible minimum heuristic algorithm have been designed, according to the minimum assignment of determining the repairing task variable of repairing resource contention, again according to repairing task resource demand, according to the minimum repairing task Resources allocation of giving of resource flexibility, determine the assignment of recombine resource variable.

(7) after definite recombine resource variable assignments, require respectively according to begin simultaneously, finish simultaneously, complete on time according to the decision-maker respectively, three kinds of time assignment strategies carry out assignment for respectively task replanning time and resource restructuring time.

Advantage of the present invention is: the method by variable dimensionality reduction, classification and substep assignment can be found the solution multitask, many resources, the concurrent repairing mission requirements of constrained generality resource restructuring decision-making, and applicability is strong, computing velocity is fast, computational accuracy is high.The present invention is directed to the characteristics of general BDAR both resource restructuring decision-making, the thought of utilizing Constraint-based to propagate, by repairing resource requirement constraint and repairing time-constrain, consistance to repairing task and recombine resource detects, when truly having inconsistent variable, problem can effectively to the variable dimensionality reduction, realize reducing the purpose of computing scale.according to the resource requirement conflict of different repairing tasks in the repairing time, BDAR both resource restructuring decision-making is divided into the repairing resource restructuring that do not conflict again on this basis, conflict repairing resource restructuring and not exclusively repairing resource restructuring fully, on this basis for Question Classification, the heuristic operators of definite type such as and resource flexibility minimum minimum based on the total resources consumption rate have been designed respectively, and mix the part classifying problem is found the solution by rounding particle cluster algorithm with binary, pointed classified calculating method is when guaranteeing computational accuracy and speed, can satisfy the demand of modern military joint protection, have more application value in army.

Embodiment

For better explanation technical scheme of the present invention, below embodiments of the present invention are described further.

(1) initialization of BDAR both resource restructuring decision-making.Suppose that BRT is n dimension repairing assignment decisions variable, RRS is v * n * m dimension restructuring repairing resource decision variable, and RRT is the n denapon variable of TT, and RS is v * m dimension repairing resource vector, and TT is the n * m denapon variable of RRS.N represents the quantity of repairing task to be responded, and m represents resource provisioning point quantity, and v represents to rush to repair resource class quantity, and BDAR both restructuring decision problem can be described as RDP={BRT so, RRS, and RRT, RS, TT, C}, model can be described as

$\max :\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{BRt}}_i\×{\mathrm{RTb}}_i---\left(1\right)$

st：BRt _i{0，1}；

BRt _i∈{BRt ₁，BRt ₂，…，BRt _n}；

RTb _i ∈{RTb ₁，RTb ₂，…，RTb _n}；

${\mathrm{RRs}}_{\mathrm{ij}}^o\∈\{\mathrm{0,1,2},\·\·\·,{\mathrm{Rs}}_j^o\};$

C＝C _t∧C _r∧C _l (2)

Wherein

i＝1，2，…，n，j＝1，2，…，m，o＝1，2，…，v (3)

i＝1，2，…，n，j＝1，2，…，m，o＝1，2，…，v (4)

i＝1，2，…，n，j＝1，2，…，m，o＝1，2，…，v (5)

● BRt _iAbout C _rAlgorithm for Consistency Checking

BRt _iAbout C _rConsistency detection be mainly to detect BRt _i=1 o'clock, all kinds of the total resources whether demand of all kinds of resources of repairing task can provide greater than all resource points.Concrete steps are

Step 1 is by the available resources moment matrix

Calculate Rs ^o, wherein

Represent j resource points, the resource available quantity of o class resource, Rs ^oRepresent all resource points, the available quantity of o class resource has ${\mathrm{Rs}}^o=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{Rs}}_j^o;$

Step 2 order

I=1, o=1, wherein

Be i item repairing task, the demand of o class resource;

Step 3 relatively

And Rs ^oIf,

O=o+1, otherwise BRt _i=0, if o＜v, repeating step 3, otherwise i=i+1, if i＜n, repeating step 3;

Step 4 is deleted all BRt from variables set BRT _i=0, record BRt _i, all and BRt _iCorrelated variables all composes zero, and output BRT.Finish to calculate.

● BRt _iAbout C _tAlgorithm for Consistency Checking

BRt _iAbout C _tConsistency detection be mainly to detect BRt _i=1 o'clock, whether the resource restructuring time of repairing task satisfied.Its algorithm steps is as follows:

Step 1 makes TT=(Tt _ij) _{N * m}, RLT=(RLt _i) _{N * 1}, RT=(Rt _i) _{N * 1}, minTt=(minTt _i) _{N * 1}, Tt wherein _ijRepresent that j resource points recombinate the i item repairing required by task time, RLt _iThe task off period that represents i item repairing task, Rt _iRepresent the i item repairing required by task repairing time, minTt _iRepresent all resource points to the i item repairing task replanning required shortest time of resource, and minTt is arranged _i=min{Tt _ij, j=1,2 ..., m;

Step 2 judgement RLt _i-Rt _iIf, RLt _i-Rt _i〉=0, judgement RLt _i-Rt _i〉=minTt _iIf, be no, BRt _i=0, i=i+1, if i＜n, repeating step 2;

Step 3 is deleted all BRt from variables set BRT _i=0, record BRt _i, all and BRt _iCorrelated variables all composes 0, and output BRT.Finish to calculate.

● BRt _iAbout C _rAnd C _tAlgorithm for Consistency Checking

BRt _iAbout C _rAnd C _tAlgorithm for Consistency Checking, be mainly to detect BRt _iAt C _rAnd C _tConsistance under the binary constraint.Can namely detect the resource points that can cover in maximum allows reorganization time, all kinds of resources of repairing required by task are provided, its algorithm steps be as follows:

Step 1 order

TT=(Tt _ij) _{N * m}, RLT=(RLt _i) _{N * 1}, RT=(Rt _i) _{N * 1}, define the same;

Step 2 makes PRT=(PRt _i) _{N * 1}, PRt wherein _iRepresent the maximum reorganization time that allows of i item repairing task, PRt is arranged _i=RLt _i-Rt _i, i=1,2 ..., n;

Step 3 order

Wherein

Represent the repairing resource points that i item repairing task allows reorganization time to cover in maximum, the o class resource available quantity that can provide has ${\mathrm{Rs}}_i^o\left(\mathrm{BRT}\right)={\mathrm{\Σ}}_{j\∈({\mathrm{Tt}}_{\mathrm{ij}}\∈{\mathrm{PRt}}_i)}{\mathrm{Rs}}_j^o,$ Make i=1, o=1;

Step 4 relatively With

If O=o+1, otherwise BRt _i=0, if o＜v, repeating step 4, otherwise i=i+1, if i＜n, repeating step 4;

Step 5 is deleted all BRt from variables set BRT _i=0, record BRt _i, all and BRt _iCorrelated variables all composes 0, and output BRT.Finish to calculate.

● About C _tAlgorithm for Consistency Checking

About C _tConsistency detection be mainly to detect

The time, whether resource can be transported in reorganization time, its algorithm steps is as follows:

Step 1 makes TT=(Tt _ij) _{N * m}, RLT=(RLt _i) _{N * 1}, RT=(Rt _i) _{N * 1}, define the same, i=1, j=1;

Step 2 judgement Tt _ij≤ RLt _i-Rt _iIf, be yes, Tt _ij=1, otherwise Tt _ij=0, i=i+1, if i＜n, if repeating step 2 is Tt _ijAll equal 0, make Rs _j=0, j=j+1, if j＜m, repeating step 2, otherwise forward step 3 to;

Step 3 is all Rs of deletion from repairing resource set RS _j=0, delete from recombine resource variables set RSS ${\mathrm{RRs}}_{\mathrm{ij}}^o\{j\∈({\mathrm{Rs}}_j=0)\},$ And output RS and RSS.Finish to calculate.

(2) the BDAR both resource restructuring Decision Classfication after initialization.After initialization BDAR both resource restructuring decision-making can be divided into resource without conflict restructuring, resource is conflicted fully and the incomplete decision-making of resource.Wherein satisfy 6 formulas without the conflicted resource restructuring, expression is the repairing resource points arbitrarily, can provide required all kinds of resources to all repairing tasks that satisfies this resource points resource restructuring time requirement.

$\∀{\mathrm{Rs}}_j^{*}\⋐{\mathrm{RS}}^{*}:{\mathrm{Rs}}_j^{*}\≥{\mathrm{\Σ}}_{\{i^{*}:{\mathrm{Tt}}_{\mathrm{ij}}^{*}\≤{\mathrm{PRt}}_i\}}{\mathrm{RSs}}_i^{*}---\left(6\right)$

7 formulas are satisfied in the conflicted resource restructuring fully, are illustrated in effective reorganization time and can provide all kinds of repairing resources to all repairing tasks, but have at least a class repairing resource can not satisfy such resource requirement of all tasks fully.

Not exclusively 8 formulas are satisfied in conflicted resource restructuring decision-making, and expression has at least a class resource of a repairing resource points to satisfy the repairing task that this resource points reorganization time requires to all provides required resource.

(3) finding the solution without conflicted resource restructuring decision-making based on the prospect algorithm.The basic thought of algorithm is from first repairing resource points, and order is to maximum repairing task and the recombine resource assignment that allows reorganization time to cover of this repairing resource points.The specific algorithm step is as follows:

Step 1 reads set ${\mathrm{BRT}}^{*}=\left\{{\mathrm{BRt}}_i^{*}\right\},$ ${\mathrm{RSS}}^{*}=\left\{{\mathrm{RSs}}_i^{o*}\right\},$ ${\mathrm{RT}}^{*}=\left\{{\mathrm{Rt}}_i^{*}\right\},$ ${\mathrm{RLT}}^{*}=\left\{{\mathrm{RLt}}_i^{*}\right\},$ ${\mathrm{RSS}}^{*}=\left\{{\mathrm{RSs}}_i^{o*}\right\},$ ${\mathrm{TT}}^{*}=\left\{{\mathrm{Tt}}_{\mathrm{ij}}^{*}\right\},$ (i＝1，2，…，n ^*，j＝1，2，…，m ^*，；o＝1，2，…，v ^*)

Step 2 judgement Tt _ij≤ RLt _i-Rt _i, getting i is 1 to n ^*, and obtain the set that each repairing resource points allows can respond in reorganization time the repairing task

(p ∈ 1 ..., n ^*), p is for arranged sequentially from small to large;

Step 3 also makes j=1 get respectively

P ∈ 1 ..., n ^*, and upgrade BRT ^*, order ${\mathrm{BRT}}^{*}={\mathrm{BRT}}^{*}-\mathrm{BRt}\left({\mathrm{Rs}}_j^{*}\right),$ If

Forward step 5 to;

Step 4j=j+1, if

Order $\mathrm{BRt}\left({\mathrm{Rs}}_j^{*}\right)={\mathrm{BRT}}^{*}\∩\mathrm{BRt}\left({\mathrm{Rs}}_j^{*}\right)$ Forward step 2 to, if j＞m ^*Forward step 5 to, otherwise repeating step 4;

The assignment of the relevant decision variable of step 5 output generates repairing task assignment and recombine resource assignment decision scheme,

Finish to calculate.

(3) find the solution based on the full resource restructuring decision-making that mixes discrete binary population.Algorithm idea utilizes this operator at first to obtain the variable assignments of one group of repairing task, and variable is carried out descending sort at first designing the greedy operator based on the wastage in bulk or weight rate.When generating the primary group, this solution is implanted initial population as a particle.Occur the infeasibility solution in the iterative process of particle cluster algorithm, last begins to set to 0 from particle, and in population, all infeasibility solutions all are modified to the feasibility solution until generate.

Step 1 is calculated the gc operator according to formula 9

$\left\{\begin{array}{c}\mathrm{gc}=\frac{{\mathrm{RTb}}_i^{*}}{C_i}\\ C_i=\underset{o=1}{\overset{v}{\mathrm{\Σ}}}({\mathrm{RSs}}_i^{o*}/\underset{j=1}{\overset{m*}{\mathrm{\Σ}}}{\mathrm{Rs}}_j^{o*})\\ j=\mathrm{1,2},\·\·\·,m^{*},o=\mathrm{1,2},\·\·\·,v\end{array}\right.---\left(9\right)$

Step 2 will be rushed to repair task and be sorted according to the descending chain of gc, be denoted as

Record

And BRT ^*The corresponding situation of the subscript of each element

Order corresponding repairing task resource demand and total resources supply are respectively $\stackrel{\→}{{\mathrm{RRS}}^{*}}=\left[\begin{array}{cccc}\stackrel{\→}{{\mathrm{RRs}}_{\mathrm{1,1}}^{*}}& \stackrel{\→}{{\mathrm{RRs}}_{\mathrm{1,2}}^{*}}& \·\·\·& \stackrel{\→}{{\mathrm{RRs}}_{1,n}^{*}}\\ \stackrel{\→}{{\mathrm{RRs}}_{\mathrm{2,1}}^{*}}& \stackrel{\→}{{\mathrm{RRs}}_{\mathrm{2,2}}^{*}}& \·\·\·& \stackrel{\→}{{\mathrm{RRs}}_{2,n}^{*}}\\ \·\·\·& \·\·\·& & \·\·\·\\ \stackrel{\→}{{\mathrm{RRs}}_{v,1}^{*}}& \stackrel{\→}{{\mathrm{RRs}}_{v,2}^{*}}& \·\·\·& \stackrel{\→}{{\mathrm{RRs}}_{v,n}^{*}}\end{array}\right]$ With ${\mathrm{RS}}^{*}={\left[\begin{array}{cccc}{\mathrm{Rs}}_1^{*}& {\mathrm{Rs}}_2^{*}& \·\·\·& {\mathrm{Rs}}_v^{*}\end{array}\right]}^T,$ First will

The full assignment of element be 0;

Step 3 makes i=1,

Step 4 judgement

If

I=i+1; Otherwise

I=i+1 is as i＞n ^*The time, forward step 4 to, otherwise, repeating step 3;

Step 5 generates initial population with X _i=(x _i1, x _i2..., x _iN), and the solution of step 1～4 gained is inserted initial population as a particle, set at random the initial velocity v of each particle _id

Step 6 is carried out the validation verification of each particle in population.Order

${\mathrm{RS}}^{*}={\left[\begin{array}{cccc}{\mathrm{Rs}}^{1*}& {\mathrm{Rs}}^{2*}& \·\·\·& {\mathrm{Rs}}^{v*}\end{array}\right]}^T,$ N=n wherein ^*, m represents population in population here;

If step 7 RSS ^*X ^T≤ RS ^*, go to step 8, make RSS otherwise find out ^*X ^T＞RS ^*X _i, i ∈ m makes X _iN classify 0 as, N=N-1, if repeating step 7 is X _iBe sky, go to step 8, otherwise repeating step 7;

Step 8 output X;

Step 9 is calculated population of future generation, to infeasibility solution in population, transforms by step 6～8, and through type 10 calculates the fitness value of each particle;

Step 10 is to each particle, relatively the desired positions p that lives through of its fitness value and it _idFitness value, if compare p _idGood, upgrade p _id

Step 11 is to each particle, relatively the desired positions p that lives through of its fitness value and colony _gdFitness value, if compare p _gdGood, upgrade p _gd

Step 12 is upgraded particle rapidity and position according to formula 11;

$v_{\mathrm{id}}^{(k+1)}={\mathrm{\ωv}}_{\mathrm{id}}^k+c_{1}r{\mathrm{and}}_1(p_{\mathrm{id}}-x_{\mathrm{id}}^k)+c_2{\mathrm{rand}}_2(p_{\mathrm{gd}}-x_{\mathrm{id}}^k)$

$x_{\mathrm{id}}^{k+1}=\left\{\begin{array}{cc}1& \mathrm{rand}<\mathrm{sigmoid}\left(v_{\mathrm{id}}^k\right)\\ 0& \mathrm{rand}>\mathrm{sigmoid}\left(v_{\mathrm{id}}^k\right)\end{array}\right.---\left(11\right)$

P wherein _i=(p _i1, p _i2..., p _iN), i=1,2 ..., m is the optimal location of i particle, P _g=(p _g1, p _g2..., p _gN) be the optimal location that whole colony finds, g is iterations, ω is inertia weight; c ₁, c ₂: acceleration constant, rand ₁, rand ₂It is the random function of even variation in [0,1] scope at random.

Step 13 judgement termination condition, discontented lumping weight is multiple 2, satisfies and finishes to calculate.

(4) based on repairing resource contention and the flexible minimum resource heuristic algorithm that not exclusively conflicts.Its main thought is at first according to time-constrain, to determine that each task covers the repairing resource points from the permission reorganization time, obtainable maximum repairing stock number.Can use situation according to resource requirement situation and the repairing resource of repairing task again, determine the resource consumption rate of this repairing task, task with repairing benefit and consumption rate ratio maximum, preferential answering as the minimum repairing task of resource contention, and from the repairing resource that can respond this repairing task, the poorest flexible resource begins assignment, until satisfy repairing task resource demand.Each to before repairing task assignment, resource contention and resource flexibility are all dynamic changes.At first according to actual assignment situation, deduct recombine resource institute assignment from repairing resource points codomain, and upgrade repairing resource points available volume of resources.Secondly whether judgement repairing resource points can cover the repairing task-set, comprise to be responded the repairing task, deducts if comprise can cover from rushing to repair resource points the quantity of repairing task the quantity that is responded the repairing task.Whether judgement has the maximum repairing resource available quantity after the resource requirement of repairing task exceeds renewal after the resource points available volume of resources is upgraded, do not respond if any this task.Repeat above-mentioned steps, until all tasks are disposed.The specific algorithm step is as follows:

The repairing task of step 1 pair input and repairing resource allow the covering difference cluster of reorganization time, order according to maximum

For maximum allows can respond the repairing resource point set of this task under reorganization time, For maximum allows under reorganization time, resource points can respond the repairing task-set that arrives;

Step 2 is calculated respectively $\mathrm{RC}\left({\mathrm{RTb}}_i^{*}\right)={\mathrm{RTb}}_i^{*}/C_i,$ Wherein $C_i=\underset{o=1}{\overset{v}{\mathrm{\&Sigma;}}}({\mathrm{RSs}}_i^{o*}/\mathrm{sum}\left({\mathrm{Rs}}^{o*}\right)$ Expression resource consumption rate, $\mathrm{sum}\left({\mathrm{RS}}^{o*}\right)=\mathrm{\&Sigma;}{\mathrm{Rs}}_j^{o*}$ $(j\&Element;{\mathrm{Tt}}_{\mathrm{ij}}^{*}\&le;{\mathrm{RPt}}_i^{*}),$ As

Be that zero explanation does not have and can to the resource points of task replanning resource, make BRt _i=0, and delete BRt from task-set to be selected _iOtherwise calculate Wherein Be resource contention (the large representative conflict of numerical value is little, otherwise the representative conflict is large),

Representative repairing resource flexibility, $\mathrm{RF}\left({\mathrm{Rs}}_j^{*}\right)=\{\mathrm{RF}\left({\mathrm{Rs}}_j^{1*}\right),\mathrm{RF}\left({\mathrm{Rs}}_j^{2*}\right)$ $,\&CenterDot;\&CenterDot;\&CenterDot;,\mathrm{RF}\left({\mathrm{Rs}}_j^{v*}\right)\}$

Step 3 from

Maximum task variable begins assignment, makes BRt _i=1, $i=i\&Element;\max \mathrm{RC}\left({\mathrm{RTb}}_i^{*}\right),$ o＝1；

Step 4 order $\stackrel{\&LeftArrow;}{\mathrm{RF}\left({\mathrm{Rs}}_j^{o*}\right)}=\{\stackrel{\&LeftArrow;}{\mathrm{RF}\left({\mathrm{Rs}}_1^{o*}\right)},\stackrel{\&LeftArrow;}{\mathrm{RF}\left({\mathrm{Rs}}_2^{o*}\right)},\&CenterDot;\&CenterDot;\&CenterDot;,\stackrel{\&LeftArrow;}{\mathrm{RF}\left({\mathrm{Rs}}_m^{o*}\right)}\}$ It is the ascending chain of o class resource flexibility;

Step 5 order $\stackrel{\&LeftArrow;}{j}=1,$ If $\stackrel{\&LeftArrow;}{j}\&Element;\stackrel{\&LeftArrow;}{{\mathrm{Tt}}_{\mathrm{ij}}^{*}}\&le;{\mathrm{PRt}}_i^{*},$ And $\stackrel{\&LeftArrow;}{{\mathrm{Rs}}_j^{o*}}-{\mathrm{RSs}}_i^{o*}\&GreaterEqual;0,$ $\stackrel{\&LeftArrow;}{{\mathrm{RRs}}_{\mathrm{ij}}^{o*}}={\mathrm{RSs}}_i^{o*},$ Otherwise $\stackrel{\&LeftArrow;}{{\mathrm{RRs}}_{\mathrm{ij}}^{o*}}=\stackrel{\&LeftArrow;}{{\mathrm{Rs}}_j^{o*}},$ $\stackrel{\&LeftArrow;}{{\mathrm{Rs}}_j^{o*}}=\stackrel{\&LeftArrow;}{{\mathrm{Rs}}_j^{o*}}-\stackrel{\&LeftArrow;}{{\mathrm{RRs}}_{\mathrm{ij}}^o},$ ${\mathrm{RSs}}_i^{o*}={\mathrm{RSs}}_i^{o*}-\stackrel{\&LeftArrow;}{{\mathrm{RRs}}_{\mathrm{ij}}^{o*}}.$ If ${\mathrm{RSs}}_i^{o*}>0,$

Repeating step 5.If Record And calculate and upgrade

O=o+1, if o＜v, repeating step 4.Otherwise, upgrade

Repeating step 2.If the flexible same case of two similar resources of resource points occurs in assignment, preferential selection can respond the few resource points of task quantity, if it is also identical select arbitrarily one to respond task quantity;

If step 6 have a class resource all can only to

Compose zero, or the whole assignment of task variable, finish to calculate.Otherwise go to step 2.

(5) according to decision-maker's Location of requirement reorganization time assignment strategy.The time assignment can be respectively begins simultaneously according to rushing to repair simultaneously the resource restructuring time, and suc as formula 12, the repairing resource restructuring time finishes simultaneously, and suc as formula 13, the repairing resource restructuring is completed on time, suc as formula 14.

The main thought that the repairing resource restructuring time begins simultaneously is that after the variable that sets the tasks, resource restructuring variable assignments, all resource restructurings begin immediately.For the repairing task of any definite response, resource restructuring time, task replanning time, repairing start time, satisfied following relation of repairing concluding time

$\left\{\begin{array}{c}{\mathrm{Tt}}_{\mathrm{ij}\_\mathrm{strat}}^{*}=t\\ {\mathrm{Rt}}_{i\_\mathrm{strat}}^{*}={\mathrm{RRt}}_{i\_\mathrm{end}}^{*}=t+\max \left\{{\mathrm{Tt}}_{\mathrm{ij}}^{*}\right\},\{j^{*}:{\mathrm{Tt}}_{\mathrm{ij}}^{*}\&le;{\mathrm{PRt}}_i\}\\ {\mathrm{Rt}}_{i\_\mathrm{end}}^{*}={\mathrm{Rt}}_{i\_\mathrm{strat}}^{*}+{\mathrm{Rt}}_i^{*}\end{array}\right.---\left(12\right)$

Wherein

For each resource points to BRt _iThe start time of restructuring repairing resource, namely give the variable assignments current time.

Be BRt _iThe repairing task replanning concluding time, by current time and this restructuring task maximum resource time.

Be the repairing start time, the repairing start time is same

Be to rush to repair the task concluding time, by repairing start time and repairing time and definite.

The main thought that the repairing resource restructuring time finishes simultaneously is, after the variable that sets the tasks, resource restructuring variable assignments, not all resource restructuring all begins immediately, but the resource restructuring time the longlyest first begins, then begin successively according to from long to short order, guarantee that the repairing resource of each task arrives simultaneously.For the repairing task of any definite response, resource restructuring time, task replanning time, repairing start time, satisfied following relation of repairing concluding time:

$\left\{\begin{array}{c}{\mathrm{Tt}}_{\mathrm{ij}\_\mathrm{strat}}^{*}=t+\max \left\{{\mathrm{Tt}}_{\mathrm{ij}}\right\}-{\mathrm{Tt}}_{\mathrm{ij}}^{*},\{j^{*}:{\mathrm{Tt}}_{\mathrm{ij}}^{*}\&le;{\mathrm{PRt}}_i\}\\ {\mathrm{Rt}}_{i\_\mathrm{strat}}^{*}={\mathrm{RRt}}_{i\_\mathrm{end}}^{*}=t+\max \left\{{\mathrm{Tt}}_{\mathrm{ij}}^{*}\right\}\\ {\mathrm{Rt}}_{i\_\mathrm{end}}^{*}={\mathrm{Rt}}_{i\_\mathrm{strat}}^{*}+{\mathrm{Rt}}_i^{*}\end{array}\right.---\left(13\right)$

The repairing resource restructuring is completed main thought on time and is, all repairing resources are as long as transport to before maximum allows the task replanning time to finish on time.For the repairing task of any definite response, resource restructuring time, task replanning time, repairing start time, satisfied following relation of repairing concluding time:

$\left\{\begin{array}{c}{\mathrm{RRt}}_{i\_\mathrm{strat}}^{*}={\mathrm{RLt}}_i^{*}-{\mathrm{Rt}}_i^{*}-t-{\mathrm{Tt}}_{\mathrm{ij}}^{*},\{j^{*}:{\mathrm{Tt}}_{\mathrm{ij}}^{*}\&le;{\mathrm{PRt}}_i\}\\ {\mathrm{Rt}}_{i\_\mathrm{strat}}^{*}={\mathrm{RRt}}_{i\_\mathrm{end}}^{*}={\mathrm{RLt}}_i^{*}-{\mathrm{Rt}}_i^{-*}\\ {\mathrm{Rt}}_{i\_\mathrm{end}}^{*}={\mathrm{RLt}}_i^{*}\end{array}\right.---\left(14\right)$

Can make the decision-maker when receiving multitask resource requirement request by decision-making technique of the present invention, the variable assignments such as repairing task, recombine resource, reorganization time are optimized decision-making, generate BDAR both resource restructuring Optimal Decision-making scheme.

Claims (1)

1. BDAR both resource restructuring Study on Decision-making Method for Optimization, it is characterized in that comprising following concrete steps: (1) carries out modeling according to the constraint satisfaction modeling method to the BDAR both resource decision, model is take the BDAR both maximizing the benefits as target, constraint condition mainly comprises repairing task resource demand type and quantitative requirement, and the restriction of repairing time.(2) according to the quantity constraint of repairing resource kind and the constraint of repairing time, by the method for constraint propagation, problem is carried out initialization, check repairing task is about repairing resource and repairing time consistency.The repairing mission requirements of recombinating in the reorganization time that after initialization, each repairing task has at least one group of recombine resource variable assignments combination to limit point.Simultaneously each repairing task supply centre provide certain repairing resource of a measurement unit can at least a repairing task.(3) require in the repairing time according to different task, between the repairing task, the situation of the conflict of demand repairing resource will rush to repair the resource restructuring decision-making and will be divided into following three classes: rush to repair the resource restructuring decision-making without conflict, conflict repairing resource restructuring decision-making and conflict repairing resource restructuring decision-making fully fully, and model is set up in classification.(4) according to rush to repair resource restructuring decision model characteristics without conflict, the repairing task has at least a resource provisioning point that all kinds of resources of repairing required by task can be provided within the restriction repairing time, belongs to " drug on the market " or " balance between supply and demand ".Giving the assignment of repairing assignment decisions variable is 1, and the task of namely rushing to repair all responds.Establish on this basis two-phase problem " the repairing resource points is minimum " and " residue repairing resource flexibility is maximum ", and determine respectively the assignment of recombine resource variable by the heuristic search algorithm of correspondence.(5) according to the repairing resource restructuring decision model characteristics of conflicting fully, designed the mixture model method for solving based on resource wastage in bulk or weight rate and particle cluster algorithm, determine the assignment of repairing task variable, again according to the heuristic of residue repairing resource flexibility maximum, determine the assignment of recombine resource variable on this basis.(6) rush to repair the characteristics of resource restructuring decision model according to not exclusively conflicting, repairing resource contention and flexible minimum heuristic algorithm have been designed, according to the minimum assignment of determining the repairing task variable of repairing resource contention, again according to repairing task resource demand, according to the minimum repairing task Resources allocation of giving of resource flexibility, determine the assignment of recombine resource variable.(7) after definite recombine resource variable assignments, require respectively according to begin simultaneously, finish simultaneously, complete on time according to the decision-maker respectively, three kinds of time assignment strategies carry out assignment for respectively task replanning time and resource restructuring time.Thereby solve " restructuring to as if who " in the BDAR both resource restructuring, " whom completes restructuring by ", problems such as " when completing restructuring ".