[12]Multi-Target Optimized Decision in the IT Surveillance (English)

8/4/2019 [12]Multi-Target Optimized Decision in the IT Surveillance (English)

1/11

Multi-Target Optimized Decision in the IT Surveillance

Bi Th Khang

H Ni, Ngy 23/06/2011


2/11

INTRODUCTION

Along with the development of information engineering surveillance,

the manager must select reasonable optimized technology of network

plan "I .To make construction duration shorter, investment lower andthe scientific disposition of manpower and material resources, it must

carry out optimization of network plan.

Usually, it is often conflict between each target of information

engineering project. For instance, the engineering construction is rapid

to build but the cost is up, sometimes the engineering cost is saved, but

the project time is longer.

It is impossible that an engineering project planning is finished with

minimum construction duration, minimum resource and lowest

investment at the same time.


3/11

INTRODUCTION(continue)

This is a multi-target optimized problem. Network planned optimization

is a complex problem and the traditional methods are still not enough

to satisfied.

The traditional methods of optimization network plan are single-

target methods. Under the certain condition, we can get a satisfied

single-target of project through the continuously adjusting and seeking.

When we think about the multi-target such as the optimization ofresource, cost and engineering time, we must use the new method that is

multi-target method. It is very important to determine a relatively

satisfactory scheme in existing feasible scheme by trading off various

targets synthesize.


4/11

THE MATHEMATICS OPTIMIZATION MODEL

OF MULTI-TARGET NETWORK PLANNED

We can use T to show the set up all selecting network picture (T1,..,Tn)

in the network project. There are m target in network project, and each

target has a different meaning (such as construction engineering timelimit, engineering cost and resource etc.). There, the goal of multi-target

decision problem can be described : k : The importance weight of sub-target k.

Z kp : The goal value of the p in selected network.

Z k0 : The best value of sub-target k.

n : The selected number of network picture.

We consider three sub-target (m = 3) such minimum engineering time,

minimum cost and the minimum resource, each sub-target can be calculated

by following :

: Minimum project time for only considering first sub-target.

: Minimum cost for only considering second sub-target.

: Minimum used resource for only considering third goal.


5/11

A. Engineering time :

There two method that can reduce the critical path :

o Break down the critical work to many small work.o Organize parallel activity or parallel overlapping activity.

Example : There are two work, A work needs to be completed for 15 says, B work

need to be completed for 30 days , if the condition allows, work A and work B will

be doing at the same time, This is illustrated in figure, the engineering time has

changed from 45 to 35 days :

The engineering time is independent work.

The engineering time has changed from 45 to 35 days.

THE CALCULATION OF EACH TARGET(1)


6/11

B. Cost :

The total costs of network plan include two parts, one is direct cost and the other is

indirect cost. The relationship between the project cost as below :The directly cost rate of working i-j is expressed with Si-j

DNi-j - the normal duration of working i-j

DCi-j - the minimum duration of working i-j

CCi-j -the direct cost needed for working i-j of minimum

CNi-j - the direct cost needed for working i-j of normal

Here is a procedure for scheduling a project. In this case, we identify the activities

as follows :



7/11

C. Resource of project :

A various resources (such as labor force, mechanical and material etc.) that one

company can offer in certain time havecertain limit, this has one problem howto use these resources economy and efficiently It is known that planned

engineering duration is T, so the average requirement (Rm) of resource for T is:

(The average requirement of resource)

(The resource distribution of

enure plan unbalanced level E)

(The unbalanced level of resource)

Obviously, the resource requirement for day by day is the more changing with

the higher of E. So, we want to make resource allocation be balanced as far as

possible when the E value reach to minimum.



8/11

C. Resource of project:

Example the optimized process :



9/11

Because of the optimization network plan is a multi-target, problem, so we

need to select a satisfactory network plan in many targets in synthesize

consideration according to determine each target weight coefficient. Themeaning of weight coefficient has reflected the important level between

each target essentially. It valde range is [0,1]

The target can be compared each other to get every target weight, and than we get

the judgment matrix B as following

Each target weight coefficient ak

DETERMINE EACH TARGET WEIGHT

COEFFICIENT k


10/11

It must point out that the multi-target optimization of network

planned is a very complex problem.

This slide has made some preliminary theories research for this

problem, and given an optimization model of multi-target network

plan.

This model is a model with general comparison, and does not get the

restriction of target number, can select proper target number

according to actual condition.

CONCLUSION


11/1114/04/2012MBA Trn Vit Dng 11

Documents

[12]Multi-Target Optimized Decision in the IT Surveillance (English)