Event Classification System to Reconsider the Production Planning

Andrés Boza, Faustino Alarcón, M. M. E. Alemany and Llanos Cuenca

Research Centre on Production Management and Engineering (CIGIP),

Universitat Politècnica de València, Valencia, Spain

Keywords: Production Planning, Event Management, Decision Making, Information System.

Abstract: On-going production planning can be affected by unexpected events that modify the planned scenario.

Decision makers must react in some way to avoid problems in the production system when these events

appear. Thus providing information about these events can be critical for decision makers, as events can

affect not only operational processes but also previously made decisions. We herein propose an event

monitoring system that classifies events into different impact levels. This information helps production

decision-makers to consider changes in the on-going production planning. Furthermore, a risk matrix can be

built, which determines the significance of the risk from the impact level and the likelihood. This likelihood

of occurrences of these events can be estimated from the historical information of the event monitoring

system. A prototype has been built following this proposal.

1 INTRODUCTION

Planning deals with finding plans to achieve some

goal (Barták 1999) and production planning is a

partial planning approach for a particular function of

a company (Buzacott et al. 2012). Production

planning also usually covers the allocation of

activities to factory departments, which is a typical

scheduling task. Production planning uses

information to generate processing routes and to find

what raw material should be ordered and when

(Barták 1999).

Once production planning decisions have been

made and planning is ongoing, unexpected events

can appear. Any cause (e.g. machine breakdowns or

changes in firm orders) that endangers current

production plan validity could lead to re-generating

the entire plan (Özdamar et al. 1998). However, the

difficulties of adapting the production plans

produces that often no changes are made (Van

Wezel et al. 2006). The conception and

implementation of appropriate information and

communication systems is a basic condition for

identifying critical incidents (Shamsuzzoha et al.

2013). In this sense, Scala et al. (2013) indicate that

data collected from sensors must trigger a chain of

events leading to changes within enterprise business

process, collaboration mechanism or organizational

framework. Such changes can be achieved in terms

of simple sense-act enterprise behaviour (direct link

between sense and act) or more complex sense-plan-

act approach (decision level). Hence the first

objective of an event monitoring system is to sense

production information about a real-time

environment and to detect events.

Enterprises normally use tools that provide them

with information to make decisions. According to

(Chen 2004), Decision Support Systems (DSSs) are

designed to use decision makers' own insights and

judgments in an ad hoc, interactive analytical

modeling process, which leads to a specific decision.

So an event monitoring and management system

should interact with DSSs to manage events that

might affect previously made decisions. It should act

as a supra-system that identifies when previous

decisions are still valid or need to be reanalyzed.

Thus traditional DSS configuration should be

extended to treat event management by a monitoring

and management system, which monitors internal

and external information (Boza et al. 2014). This

event information can also be represented in the

form of rules, such as IF–THEN. For example, if a

firm order exceeds a specific value, an alert is

triggered. This set of rules represents an expert

system: it contains information obtained from a

human expert, which is represented in form of rules

(Liao 2005).

According to the ISO/Guide 73:2009 (2009), risk

is the combination of the probability of an event and

its consequences when exploiting any vulnerability.

Boza, A., Alarcón, F., Alemany, M. and Cuenca, L.

Event Classiﬁcation System to Reconsider the Production Planning.

In Proceedings of the 18th International Conference on Enterprise Information Systems (ICEIS 2016) - Volume 2, pages 82-88

ISBN: 978-989-758-187-8

So, with the information of events, risks can also be

identified. We propose herein a monitoring software

application, based on rules, that detects unexpected

events in production planning and identify risks

produced by these events. In order to explain our

purpose in this paper: Section 2 reviews problems in

production planning in the literature; Section 3 deals

with event management; Section 4 defines expert

decision support system based on the literature;

Section 5 explains our proposal to monitor and

classify events; Section 6 offers a prototype of this

proposal; Section 7 presents the conclusions drawn

from this approach.

2 INCIDENCES IN PRODUCTION

PLANNING

Production planning can be affected by different

unexpected events or incidences, for example, a

broken machine or a huge order. In the literature,

authors have dealt with these problems in different

ways. Chan et al. (2006) indicate that frequent

changes in the current schedule may lead to

disturbances in production, and may result in

lateness orders or increased production costs.

Weinstein and Chung (1999) explain that when

production equipment displays signs of failure, or

they occur after, this may adversely affect both

production plan integrity and product quality. Poon

et al. (2011) explain that in the actual manufacturing

environment, shop floor managers face numerous

unpredictable risks in day-to-day operations, such as

defects in supplies of components or raw materials,

or errors, failures and wastage in various production

processes. Baron and Paté-Cornell (1999) indicate

that during the manufacturing process, unexpected

interruptions appear, which could be accidents,

machine breakdowns or human errors. In a cookie

factory case, Van Wezel et al. (2006) study planning

flexibility and classify events according to their

source: a) Customer (e.g. rush order, change in order

volume, or earlier/later delivery date); b) Product

(e.g. raw material out of stock, too little or too much

stock of end product(s), or product sent back); c)

Process (e.g. setup/cleaning time variation,

more/less waste, or higher/ lower production speed;

d) Machines/staff (e.g. long disruptions, shortage or

surplus capacity, or variation in run-in times). All

these planning problems need to be managed and it

is necessary to decide how to deal with these events.

To deal with the events, their detection is a very

important task. Once an event occurs in a company,

event information is stored in the system and

analysis information is delivered. With this

information, decision-makers decide what action

must to take to solve the problem. If the detection of

the event is slow, the troubles will be bigger. In this

sense, new technologies like Internet of Things can

help in this purpose. A quicker identification of

relevant events is necessary to make a quicker

analysis of their consequences. SAP (2014)

highlights how value diminishes as time elapses

between when data is first captured and when an

action or decision is triggered. This analysis must

include not only a short-term point of view, but also

the consequences for the on-going production

planning.

3 EVENT MANAGEMENT

Event management consists in anticipating and

planning solutions for business events before they

appear rather than reacting after damage is

produced. The literature includes various authors

who deal with event management not only for a

company but also for business networks, such as

Virtual Organizations (Carneiro 2013) or

Collaborative Networks (Vargas et al. 2014). Baron

and Paré-Cornell (1999) provide an analytical and

dynamic link between the Risk Management System

and the long-term productivity and safety

performance of the physical system. Barash et al.

(2007) purpose a decision support tool for the

business impact analysis and improvement of the

incident management process in IT support

organization. Bartolini et al. (2010) present an

approach to assess and improve the performance of

an IT support organization in managing service

incidents based on the definition of a set of

performance metrics and a methodology. This

guided analysis allows users to find the root causes

of poor performance and to decide about the

corrective actions to be taken. Liu et al. (2007)

develop an approach for modeling event

relationships in a supply chain through Petri nets as

a formalism for managing events. Söderholm (2008)

aims to outline different categories of unexpected

events that appear in projects as a result of

environmental impacts and how these are dealt with.

Bearzotti et al. (2012) present an agent-based

approach for the Supply Chain Event Management

problem, which can perform autonomous corrective

control actions to minimize the effect of deviations

in the plan currently underway. These approaches

made different event classification to manage the

Event Classiﬁcation System to Reconsider the Production Planning

unexpected events: according to its impact (Baron

and Pate-Cornell 1999), according to its supporting

(Bartolini et al. 2010)and according to specific

groups given by the company (Liu et al. 2007;

Bearzotti et al. 2012). Only one of these research

made a monitoring system to detect events

(Bearzotti et al. 2012). But all these approaches

require an expert engineer to define the rules.

A very accepted classification of events is

according to their impact in the organization on a

scale from 1 to 5, where 1 represents the least level

and 5 the strongest (Shamsuzzoha et al. 2013).

Knowing the severity of the event, risk can be

identified by the occurrence likelihood of this event.

Thus a risk matrix can be used to classify events.

This matrix has several categories, “probability,”

“likelihood” or “frequency”, for its columns and

several categories, “severity,” “impact” or

“consequences”, for its rows. It associates a

recommended level of risk, urgency, priority or

management action with each row-column pair; that

is, with each cell (Cox Jr. 2008).

These risk matrices have been widely praised

and adopted as simple effective approaches to risk

management. According to Cox (2008), their main

advantages are that they provide: (1) a clear

framework for the systematic review of individual

risks and portfolios of risks; (2) convenient

documentation for the rationale of risk rankings and

priority setting; (3) relatively simple inputs and

outputs, often with attractively colored grids; (4)

opportunities for many stakeholders to participate in

customizing category definitions and action levels;

(5) opportunities for consultants to train different

parts of organizations on “risk culture” concepts at

different levels of detail. So the risk matrix is an

appropriate tool to classify events.

4 EXPERT DECISION SUPPORT

SYSTEM

DSSs are normally used as a tool to make decisions

when faced with certain problems. They are defined

as computer systems that deal with a problem where

at least some stage is semi-structured or

unstructured. A computer system can be developed

to deal with the structured portion of a DSS

problem, but decision makers’ judgment must

consider the unstructured part, to hence constitute a

human-machine problem-solving system (Shim et al.

2002). The primary purpose of DSSs is to help

decision-makers develop an understanding of the ill-

structured complex environment represented by the

model (Steiger 1998).

Turban and Watkins (1986) described the Expert

System like a computer program, which includes a

knowledge base that contains an expert's knowledge

for a particular problem domain, and a reasoning

mechanism for propagating inferences on the

knowledge base. The benefits generated by expert

systems include (Cohen and Asín 2001): (1) less

dependence on key personal; (2) facilitating staff

training; (3) improving the quality and efficiency of

decision making; (4) transferring the ability of

making decisions. Integrating an Expert System into

DSSs helps obtain more benefits. These benefits can

be used in several dimensions (Turban and Watkins

1986): Expert Systems contribution, DSS

contribution, and the synergy resulting from the

DSS/ES combination.

5 PROPOSAL OF AN EVENT

MONITORING SYSTEM FOR

CLASSIFICATING EVENTS TO

RECONSIDER THE

PRODUCTION PLANNING

An event monitoring application should interact with

the DSS used in the production planning system by

decision makers. However, expert knowledge is

necessary to identify and classify potential events by

their impact level. Given the advantages of the

Expert DSS presented in the previous section, we

propose an Event Monitoring System (EMS) based on

an Expert DSS, which identifies and classifies events

(CE) that have an impact on on-going production

planning and interact with the DSS used in production

planning (PP) systems, dubbed as EMS-CE-PP.

Depending on its likelihood and impact level, the

system indicates the seriousness of the event in the

previously shown standard risk matrix. This

likelihood can be estimated by the system, counting

the number of times that an event appears.

The proposed expert DSS does not use an Expert

System like an intelligent program, which

automatically makes a decision, but uses it like a

support system for decision makers.

5.1 Event Monitoring System (EMS)

Framework

Some enterprises generate their production planning

with DSSs that use mathematical models (Model-

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

Driven DSS). The decisions made with these Model-

Driven DSSs can be affected by different events. A

significant set of events to be identified includes

those that affect the planning generated by these

Model-Driven DSSs. The mathematical models used

in these DSSs are written in modeling languages,

such as Modeling Programming Language (MPL).

So it is possible to extract parameters and decision

variables from these models that can be affected by

events. The parameters and decision variables form

a set of attributes of the models.

This is the starting point for our proposal, where

an expert in production planning systems selects the

set of attributes that require a control. These

attributes will be used to make rules. A rule is a

condition defined by the decision maker to identify

the events: if this condition goes into effect, an event

alert appears. These rules are made by the expert, a

person with high knowledge about event detection in

production. This expert is usually the decision

maker.

The objective of these rules is to identify changes

in the production system to reconsider the current

production planning generated by the DSSs between

each re-planning period. The current information

about the production systems can be significantly

different from the previous information used by the

DSSs when the current planning was generated.

This proposal extends the DSS proposed by Boza

et al. (2014), which includes three phases: (1) model

and attributes selection: experts select decision

models and the attributes (of these models) that can

be affected by events; (2) criteria creation and

visualization: experts create alert criteria about

previously selected attributes; (3) execution:

validation of the alert criteria conditions executed

manually or automatically. Our proposal herein

intends to extend the previous proposal to include

the event classification and risk identification based

on the risk matrix. This information allows the

decision-maker reconsider the current production

planning. The following paragraphs review these

phases and detail our proposal.

5.2 Model and Attributes Selection

An expert in production planning systems selects the

mathematical models used in the planning

production decision system to analyze the alert

criteria on them. After selecting the models, experts

can identify the model’s parameters and decision

variables to create the alert criteria to identify

events. These selected attributes must have impact

into the production planning and its variation can

produce a modification in the production decisions.

For example: variation in demand or machine setup

times.

5.3 Criteria Creation and Visualization

Alert criteria can be defined according to the

selected attributes and a classification of the events

can be made. We propose using five impact levels

for each criterion: Extremely Serious Level, Serious

Level, Substantial Level, Moderate Level and Low

Level. Each level is achieved according to a logical

operation formed by constants, attributes and

functions. Alerts are triggered when a true value

appears in these logical operations. Constants are

values that are introduced directly by the expert;

attributes are the previously selected parameters and

decision variables; functions are operations formed

by attributes and constants, such as addition,

averages, etc.

Enterprise information is dynamic, so any

unexpected development of an attribute should be

analyzed. In order to consider this development in

the alert criteria, it is necessary the current and/or

previous values for each attribute in the alert criteria;

i.e., attributes values are taken from the current

production system state and/or from the previous

state (when the production planning was made).

Thus, decision makers introduce rules (using logical

conditions) to identify events. Table 1 shows

combinations in these logical conditions (A -logical

condition- B).

Table 1: Possible combinations of logical operation to

criteria creation.

Logical

condition

Current Attribute

Value

Attribute Value

Current or Previous

Attribute Value

Constant Value

Current or Previous

Attribute Value

Function Result

Function Result Constant Value

Function 1 Result 1 Function 2 Result

Alert criteria can also be defined for particular

objects (e.g. the demand limit value of a specific

product), or from a general perspective, (e.g. the

demand limit value of all the products).

5.4 Execution

After creating the alert criteria, decision makers can

use the Event Monitoring Systems to evaluate the

Event Classiﬁcation System to Reconsider the Production Planning

situation with these criteria. This evaluation can be

made automatically (e.g. by time intervals: hourly,

daily or weekly) or manually. During these

evaluations, the EMS-CE-PP checks the criteria

(using the rules previously introduced) with the

enterprise information, and as a result, events can be

detected and decisions makers are alerted.

5.5 Event Impact Classification

Decision makers obtain new information after each

execution. This information shows detected events

related with each criterion and the impact level that

produces that event. Also, the information about the

number of occurrences of the event is stored to have

historical information in order to obtain the

likelihood and calculate the risk.

The impact of the event had been indicated

previously by the expert and the likelihood is

estimated by the system with the information of

previous executions. This information allows

decision makers to identify the impact of the event

in order to evaluate the situation, try to solve the

problem and, if necessary, change the on-going

production planning, and to obtain information about

the event risks.

Figure 1 shows the event monitoring system

framework proposal.

6 EMS-CE-PP PROTOTYPE

An Event Monitoring System prototype to Classify

Events to reconsider the Production Planning was

developed using Java libraries. The main elements

used in the application were:

• Mathematical models used for the DSS to

propose the production planning. The

mathematical models have been defined in

Mathematical Modeling Language (MPL).

• Databases with information about production.

These databases include information about the

current situation of the production system and

the previous information of the production

system when the DSS proposed the production

planning.

• An internal database which includes the

knowledge database.

The internal database has four main tables:

attributes table to save the attributes of the model

selected by the user; a criteria table, which stores

the criteria created by decision makers; an execution

criteria table, which saves information on execution

(if execution is automatic or manually, interval time,

etc.). Once execution has been run, the results are

saved in the results table, which saves the

information on each alert criterion (attributes values,

event significance, event frequency, etc.).

6.1 Model and Attributes Selection

The EMS prototype allows the user to select an MPL

file in order to load the attributes used in this model.

An expert can select between these attributes, which

will be used to create the alert criteria. Furthermore,

a link must be created between the attribute and

Figure 1: The event monitoring system framework.

Decision

maker

DSS

Production System

Decisions

Mathematical model

- Parameters

- Decision variables

Production

information

Rules

DSS

EMS

Production

information

Information

to make

decisions

Event

Attributes

Expert

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

the database (table and column) that contain their

values.

The criteria creation form includes name, criteria

operands, the logic operation to be performed with

these operands, the impact level and a description.

Also, some attribute characteristics need to be

identified: (1) the attributes data in the criteria can

be obtained from current values or previous values;

(2) the alert criteria is general or for a particular

object. This information is stored in the internal

database.

6.2 Execution

Periodical or manual monitoring can be made using

the EMS-CE-PP prototype. The event monitoring

system obtains information from the production

databases in order to evaluate the criterion

previously defined. This evaluation of each criterion

allows identifying the impact levels for each

criterion: Extremely Serious Level, Serious Level,

Substantial Level, Moderate Level and Low Level.

If an alarm appears in several levels for the same

criterion, it is stored the most serious level.

6.3 Event Impact Classification

The information is presented like a criterion list. A

warning icon appears and indicates that an alarm

occurs in this criterion. Production information is

shown in green, yellow or red according to the

impact level. This information can be evaluated for

the decision-makers to reconsider the validity of the

current production planning.

7 CONCLUSIONS

Production processes include quite complex decision

making, and consequently, production plans are

launched. Unexpected events can appear while these

plans are on-going, which could have a major or

minor impact on these on-going plans. If the impact

is major, it can force a change to be made in the

established planning. This research proposes an

Event Monitoring Software Application based on an

expert system to identify the events and to classify

them according to their impact level on production

planning.

Experts can create production system alert

criteria. In this way, decision makers can monitor

these events and check if there are any unexpected

events.

This proposal presents some advantages: i) own

creation of impact criteria (rules) according to each

production system to classify events; ii) connection

with the DSS models used in the production

planning and the production information system; iii),

information to alert decision makers to decide

whether to change production plans or not.

Some future research lines of this system are:

• As mentioned earlier, the EMS-CE-PP alerts

the decision makers about the critical events in

their production system. However, it does not

evaluate their economical impact. Our proposal

could be improved if a cost/benefits analysis is

included in order to provide further

information.

• Most organizations have a hierarchical

structure and make hierarchical decisions. The

system could consider different hierarchical

levels in production planning and define

different sets of criteria at each planning

system level.

• Using IoT technologies: IoT technologies are

able to give further information about the

Production Systems. An Event Monitoring

System could take these technologies into

account in order to identify quickly relevant

events in the Production System and to extend

the EMS analysis with new information

gathered with these technologies.

ACKNOWLEDGEMENTS

This research has been carried out in the framework

of the project GV/2014/010 funded by the

Generalitat Valenciana (Identificación de la

información proporcionada por los nuevos sistemas

de detección accesibles mediante internet en el

ámbito de las “sensing enterprises” para la mejora de

la toma de decisiones en la planificación de la

producción).

REFERENCES

Barash, G., Bartolini, C., & Wu, L. 2007, May. Measuring

and improving the performance of an IT support

organization in managing service incidents. In

Business-Driven IT Management, 2007. BDIM'07. 2nd

IEEE/IFIP International Workshop on (pp. 11-18).

IEEE.

Baron, M. M., & Pate-Cornell, M. E. 1999. Designing

risk-management strategies for critical engineering

Event Classiﬁcation System to Reconsider the Production Planning

systems. Engineering Management, IEEE

Transactions on, 46(1), 87-100.

Barták, R. 1999, December. On the boundary of planning

and scheduling: a study. In Proceedings of the

Eighteenth Workshop of the UK Planning and

Scheduling Special Interest Group (pp. 28-39).

Bartolini, C., Stefanelli, C., & Tortonesi, M. 2010.

Analysis and performance improvement of the IT

incident management process. Proc. Transactions on

Network and Service Management (TNSM 10), IEEE

Press, 132-144.

Bearzotti, L. A., Salomone, E., & Chiotti, O. J. 2012. An

autonomous multi-agent approach to supply chain

event management. International Journal of

Production Economics, 135(1), 468-478.

Boza, A., Cortes, B., Alemany, M. D. M. E., & Vicens, E.

2015. Event Monitoring Software Application for

Production Planning Systems. In Enhancing Synergies

in a Collaborative Environment (pp. 123-130).

Springer International Publishing.

Buzacott, J. A., Corsten, H., Gössinger, R., & Schneider,

H. M. 2012. Production planning and control: basics

and concepts. Oldenbourg Verlag.

Carneiro, L. M., Cunha, P., Ferreira, P. S., &

Shamsuzzoha, A. 2013. Conceptual framework for

non-hierarchical business networks for complex

products design and manufacturing. Procedia CIRP, 7,

61-66.

Chan, F. T., Au, K. C., & Chan, P. L. Y. 2006. A decision

support system for production scheduling in an ion

plating cell. Expert Systems with Applications, 30(4),

727-738.

Chen, K. C. 2004. Decision support system for tourism

development: system dynamics approach. The Journal

of Computer Information Systems, 45(1), 104.

Cohen D, Asín E 2001. Sistemas de información para los

negocios: un enfoque de toma de decisiones. McGraw-

Hill.

Cox Jr. LA (Tony) 2008. What’s Wrong with Risk

Matrices? Risk Analysis, 28 (2), 497–512.

ISO 2009. 73: 2009: Risk management vocabulary.

Liao, S. H. 2005. Expert system methodologies and

applications—a decade review from 1995 to 2004.

Expert systems with applications, 28(1), 93-103.

Liu, R., Kumar, A., & Van Der Aalst, W. 2007. A formal

modeling approach for supply chain event

management. Decision Support Systems, 43(3), 761-

778.

Özdamar, L., Bozyel, M. A., & Birbil, S. I. 1998. A

hierarchical decision support system for production

planning (with case study). European Journal of

Operational Research, 104(3), 403-422.

Poon, T. C., Choy, K. L., Chan, F. T., & Lau, H. C. 2011.

A real-time production operations decision support

system for solving stochastic production material

demand problems. Expert Systems with Applications,

38(5), 4829-4838.

Sacala, I., Moisescu, M., & Repta, D. 2013. Towards the

development of the future internet based enterprise in

the context of cyber-physical systems. In 2013 19th

International conference on control systems and

computer science (CSCS), 405–412.

SAP AG 2014. Next-Generation Business and the Internet

of Things. Studio SAP | 27484enUS (14/03).

Shamsuzzoha, A. H. M., Rintala, S., Cunha, P. F.,

Ferreira, P. S., Kankaanpää, T., & Maia Carneiro, L.

2013. Event Monitoring and Management Process in a

NonHierarchical Business Network. Intelligent Non-

hierarchical Manufacturing Networks, 349-374.

Shim, J. P., Warkentin, M., Courtney, J. F., Power, D. J.,

Sharda, R., & Carlsson, C. 2002. Past, present, and

future of decision support technology. Decision

support systems, 33(2), 111-126.

Söderholm, A. 2008. Project management of unexpected

events. International Journal of Project Management,

26(1), 80-86.

Steiger, D. M. 1998. Enhancing user understanding in a

decision support system: a theoretical basis and

framework. Journal of Management Information

Systems, 199-220.

Turban, E., & Watkins, P. R. 1986. Integrating expert

systems and decision support systems. Mis Quarterly,

121-136.

Van Wezel, W., Van Donk, D. P., & Gaalman, G. 2006.

The planning flexibility bottleneck in food processing

industries. Journal of Operations Management, 24(3),

287-300.

Vargas, A., Cuenca, L., Boza, A., Sacala, I., & Moisescu,

M. 2016. Towards the development of the framework

for inter sensing enterprise architecture. Journal of

Intelligent Manufacturing, 27(1), 55-72.

Weinstein, L., & Chung, C. H. 1999. Integrating

maintenance and production decisions in a hierarchical

production planning environment. Computers &

operations research, 26(10), 1059-1074.

ICEIS 2016 - 18th International Conference on Enterprise Information Systems