Eusflat2007 - Fuzzy Logic

european blog for fuzzy logic and technology

Scope of artificial Intelligence in Business



Introduction



Business applications utilize the specific technologies mentioned earlier to try and make better sense of potentially enormous variability (for example, unknown patterns/relationships in sales data, customer buying habits, and so on). However, within the corporate world, AI is widely used for complex problem-solving and decision-support techniques in real-time business applications. The business applicability of AI techniques is spread across functions ranging from finance management to forecasting and production.

In the fiercely competitive and dynamic market scenario, decision-making has become fairly complex and latency is inherent in many processes. In addition, the amount of data to be analyzed has increased substantially. AI technologies help enterprises reduce latency in making business decisions, minimize fraud and enhance revenue opportunities.



Definition of AI



AI is a broad discipline that promises to simulate numerous innate human skills such as automatic programming, case-based reasoning, neural networks, decision-making, expert systems, natural language processing, pattern recognition and speech recognition etc. AI technologies bring more complex data-analysis features to existing applications.

There are many definitions that attempt to explain what Artificial Intelligence (AI) is. I like to think of AI as a science that investigates knowledge and intelligence, possibly the intelligent application of knowledge. Knowledge and Intelligence are as fundamental as the universe within which they exist, it may turn out that they are more fundamental.

One of the aims of AI is said to be the investigation of human cognition and AI is part of Cognitive Science. AI is really an investigation into the creation of intelligence and that there is no reason for the intelligence that is created to be exactly the same as human intelligence.



Importance of AI



Enterprises that utilize AI-enhanced applications are expected to become more diverse, as the needs for the ability to analyze data across multiple variables, fraud detection and customer relationship management emerge as key business drivers to gain competitive advantage.

Artificial Intelligence is a branch of Science which deals with helping machines, finds solutions to complex problems in a more human-like fashion. This generally involves borrowing characteristics from human intelligence, and applying them as algorithms in a computer friendly way. A more or less flexible or efficient approach can be taken depending on the requirements established, which influences how artificial the intelligent behavior appears.

AI is generally associated with Computer Science, but it has many important links with other fields such as Maths, Psychology, Cognition, Biology and Philosophy, among many others. Our ability to combine knowledge from all these fields will ultimately benefit our progress in the quest of creating an intelligent artificial being.



Emergence of AI in business



Artificial Intelligence (AI) has been used in business applications since the early eighties. As with all technologies, AI initially generated much interest, but failed to live up to the hype. However, with the advent of web-enabled infrastructure and rapid strides made by the AI development community, the application of AI techniques in real-time business applications has picked up substantially in the recent past.

Computers are fundamentally well suited to performing mechanical computations, using fixed programmed rules. This allows artificial machines to perform simple monotonous tasks efficiently and reliably, which humans are ill-suited to. For more complex problems, things get more difficult… Unlike humans, computers have trouble understanding specific situations, and adapting to new situations. Artificial Intelligence aims to improve machine behavior in tackling such complex tasks.

Together with this, much of AI research is allowing us to understand our intelligent behavior. Humans have an interesting approach to problem-solving, based on abstract thought, high-level deliberative reasoning and pattern recognition. Artificial Intelligence can help us understand this process by recreating it, then potentially enabling us to enhance it beyond our current capabilities.



Applications of AI



The potential applications of Artificial Intelligence are abundant. They stretch from the military for autonomous control and target identification, to the entertainment industry for computer games and robotic pets, to the big establishments dealing with huge amounts of information such as hospitals, banks and insurances, we can also use AI to predict customer behavior and detect trends.

AI is a broad discipline that promises to simulate numerous innate human skills such as automatic programming, case-based reasoning, decision-making, expert systems, natural language processing, pattern recognition and speech recognition etc. AI technologies bring more complex data-analysis features to existing applications.

Business applications utilize the specific technologies mentioned earlier to try and make better sense of potentially enormous variability (for example, unknown patterns/relationships in sales data, customer buying habits, and so on). However, within the corporate world, AI is widely used for complex problem-solving and decision-support techniques in real-time business applications. The business applicability of AI techniques is spread across functions ranging from finance management to forecasting and product



Artificial Neural Networks



An artificial neural network (ANN), often just called a “neural network” (NN), is a mathematical model or computational model based on biological neural networks. It consists of an interconnected group of artificial neurons and processes information using a connectionist approach to computation. In most cases an ANN is an adaptive system that changes its structure based on external or internal information that flows through the network during the learning phase. In more practical terms neural networks are non-linear statistical data modeling tools. They can be used to model complex relationships between inputs and outputs or to find patterns in data.



Real life applications of ANN



The tasks to which artificial neural networks are applied tend to fall within the following broad categories:

• Function approximation, or regression analysis, including time series prediction and modeling.

• Classification, including pattern and sequence recognition, novelty detection and sequential decision making.

• Data processing, including filtering, clustering, blind source separation and compression.

Application areas include system identification and control (vehicle control, process control), game-playing and decision making (backgammon, chess, racing), pattern recognition (radar systems, face identification, object recognition and more), sequence recognition (gesture, speech, handwritten text recognition), medical diagnosis, financial applications (automated trading systems), data mining (or knowledge discovery in databases, “KDD”), visualization and e-mail spam filtering.

The proven success of Artificial Neural Networks (ANN) and expert systems has helped AI gain widespread adoption in enterprise business applications. In some instances, such as fraud detection, the use of AI has already become the most preferred method. In addition, neural networks have become a well-established technique for pattern recognition, particularly of images, data streams and complex data sources and, in turn, have emerged as a modeling backbone for a majority of data-mining tools available in the market. Some of the key business applications of AI/ANN include fraud detection, cross-selling, customer relationship management analytics, demand prediction, failure prediction, and non-linear control.

A majority of the enterprises adopt horizontal or vertical solutions that embed neural networks such as insurance risk assessment or fraud-detection tools, or data-mining tools that include neural networks (for instance, from SAS, IBM and SPSS) as one of the modeling options.

Artificial Intelligence in Manufacturing

As the manufacturing industry becomes increasingly competitive, sophisticated technology has emerged to improve productivity. Artificial Intelligence in manufacturing can be applied to a variety of systems. It can recognize patterns, plus perform time consuming and mentally challenging tasks. Artificial Intelligence can optimize your production schedule and production runs. In order for organizations to meet ever increasing customer demands, and to be able to survive in an environment where change is inevitable, it is crucial that they offer more reliable delivery dates and control their costs by analyzing them on a continual basis. For businesses, being capable of delivering high quality goods at low costs and short delivery times is akin to operating in a whirlpool environment like the Devil’s Triangle, and this is no easy task for any organization. Managing so that production takes place at the right time, on the right equipment, and using the right tools will minimize any deviations in delivery dates promised to the customer. Utilizing equipment, personnel and tools to their maximal efficiency will no doubt improve any organization’s competitive strength. In return, proper utilization of these capabilities will result in lower costs for the organization

Optimal scheduling of jobs on equipment, without the use of computer software, is a truly difficult undertaking. Performing planning using the “Deterministic Simulation Method” will provide you with schedules that will indicate job loads per equipment. Even in the case limited to a single piece of equipment, as the number of jobs to schedule on that equipment increases, finding the right solution in the “Possible Solutions Set” becomes next to impossible. And in the real world, the difficulties arising from the large size of the solutions set due to the recipes formed by jobs, equipment and products, and shaped by the technological restrictions, as well as the complexity in finding a close to ideal solution, are readily apparent.

Research and studies are being conducted worldwide on the subject of scheduling. Software vendors working in this area follow developments closely, and they are coming out with new products to better meet demands. “Genetic Algorithms”, “Artificial Intelligence”, and “Neural Networks” are some of the technologies being used for scheduling



Advantages



• View your best product runs and the corresponding settings.

• Increase efficiency and quality by using optimal settings from past production.

• Artificial Intelligence can optimize your schedule beyond normal human capabilities.

• Increase productivity by eliminating downtime due to unpredictable changes in the schedule.



Artificial Intelligence in Financial services



AI has found a home in financial services and is recognized as a valuable addition to numerous business applications. Sophisticated technologies encompassing neural networks and business rules along with AI-based techniques are yielding positive results in transaction-oriented scenarios for financial services. AI has been widely adopted in such areas of risk management, compliance, and securities trading and monitoring, with an extension into customer relationship management (CRM). Tangible benefits of AI adoption include reduced risk of fraud, increased revenues from existing customers due to newer opportunities, avoidance of fines stemming from non-compliance and averted securities trade exceptions that could result in delayed settlement, if not detected.

Warren Buffet is known as the ultimate investor in this age. So good is he, in fact, that artificial intelligence software developed in Carnegie Mellon that predicts stock movements was named after him by. But can machines really take the place of human traders, much less surpass them? When Deep Blue defeated Chess Grandmaster Kasparov in 1997, AI was propelled into the limelight. Indeed, if a machine can whiz through the intricacies of the ultimate game of strategy, why not beat man in other fields as well – thereby facilitating work, decreasing costs and errors and increasing productivity and quality. This study focuses on applying AI in Finance, particularly in stock trading. In the field of Finance, artificial intelligence has long been used. Some applications of Artificial Intelligence are

• Credit authorization screening

• Mortgage risk assessment

• Project management and bidding strategy

• Financial and economic forecasting

• Risk rating of exchange-traded, fixed income investments

• Detection of regularities in security price movements

• Prediction of default and bankruptcy

• Security/and or Asset Portfolio Management

Artificial intelligence types used in finance include neural networks, fuzzy logic, genetic algorithms, expert systems and intelligent agents. They are often used in combination with each other. When AI first appeared a decade ago, it generated mass media hype but delivered inconsistent results. A number of those who praised its ability were paralyzed in the end. One such case is Fidelity Investments. In this paper, we set the stage by describing how traditional stock trading differs from AI-powered stock trading. We define the various AI systems available and also explore the various solutions available in the market, their IT foundations and how salient they are. Then, we move into how AI systems for stock trading will affect traders, companies and individuals. Benefits, risks and competitive strategy will be defined and real-world examples cited, as grounding for our recommendations in the end. Recommendations include getting management buy-in, implementing the system and managing the whole structure to succeed.



Artificial Intelligence in Marketing



Advances in artificial intelligence (AI) eventually could turbo-boost customer analytics to give companies speedier insights into individual buying patterns and a host of other consumer habits.

Artificial intelligence functions are made possible by computerized neural networks that simulate the same types of connections that are made in the human brain to generate thought. Currently, the technology is used mostly to analyze data for genetics, pharmaceutical and other scientific research. It’s seeing little use in CRM right now, though it has tremendous potential in the future

AI-enhanced analytics programs also provide survival modeling capabilities — suggesting changes to products based on use. For example, customer patterns are analyzed to learn ways to extend the life of light bulbs or to help decide the correct dosage for medications.

High-tech data mining can give companies a precise view of how particular segments of the customer base react to a product or service and propose changes consistent with those findings. In addition to further exploring customers” buying patterns, analytics could help companies react much more quickly to the marketplace.

According to Meta Group vice president Liz Shahnam, intelligent agents could let companies make real-time changes to marketing campaigns. “New technologies would have the model refreshed on the fly based on each new incoming piece of customer information — reaction to the campaign — for a more targeted offer,”



Artificial Intelligence in HR



It is widely believed that the role of managers is becoming a key determinant for enterprises’ competitiveness in today’s knowledge economy era. Owing to fast development of information technologies (ITs), corporations are employed to enhance the capability of human resource management, which is called human resource information system (HRIS). Recently, due to promising results of artificial neural networks (ANNs) and fuzzy theory in engineering, they have also become candidates for HRIS. The artificial intelligence (AT) field can play a role in this, especially; in assuring that the fuzzy neural network has the characteristics and functions of training, learning, and simulation to make an optimal and accurate judgment according to the human thinking model. The main purposes of the study are to discuss the appointment of managers in enterprises through fuzzy neural network, to construct a new model for evaluation of managerial talent, and accordingly to develop a decision support system in human resource selection. Therefore, the research methods of reviewing literature, in-depth interview, questionnaire survey, and fuzzy neural network are used in the study. The fuzzy neural network is used to train the concrete database, based on 191 questionnaires from experts, for getting the best network model in different training conditions. In order to let decision-makers adjust weighted values and obtain decisive results of each phase’s scores, we adopted the simple additive weighting (SAW) and fuzzy analytic hierarchy process (FAHP) methods in the study. Finally, the human resource selection system of Java user interface has been constructed by FNN in the study.



Conclusion



It is difficult for business to see general relevance from AI. This is probably one of the reasons for the compartmentalization of AI into things like Knowledge Based Systems, Neural Networks, and Genetic Algorithms etc. Some of these separate sub topics have been shown to be very useful in solving certain difficult business and industrial problems and consequently funding bodies influence research directions by encouraging work on these more application based areas. This can have a positive effect for business benefit and has lead to some very useful systems that have found their way into the heart of business activity. Business should not lose sight of where AI could go because there are many potential benefits to current and new businesses of future research. The idea of robotic domestic workers is still far fetched but companies are making progress even here. There is already a Robot Vacuum Cleaner marketed by Electrolux and doubtless improved systems with better functionality will follow. .

I would like to close by quoting from Tom Peters, a leading management guru: “When you think you’ve reached the top, tear down everything and do it all over again. If you don’t, your competitor will.” To this, I would like to add my own: “If your competitor won’t, new investors will enter the market segment who will do the same job better.”



By: Sabarirajan.A

About the Author:

Larry

Key No. 3 - 7 essential criteria for selecting your ERP solution and technology partner

Once you’ve made your decision as to why you are considering an ERP implementation (covered in article #1 in this series) and investigated the total cost of ownership (article #2), there are several aspects you should consider in detail when selecting a specific system for your situation.

The seven most important of these are

• Functional compatibility with current and future business requirements

• Total cost of ownership

• Operational Metrics

• Flexibility

• Time and ease of implementation

• Vendor support and relationships

• Industry expertise and customer references

A survey by the Aberdeen Group (June 2007) found when it asked respondents what criteria were most important in selecting an ERP vendor, “remarkably little variation was visible across company size … functionality is the clear top priority for all companies, followed by total cost of ownership”.

1. Functional compatibility

The first question you need to ask is: what applications can accommodate your business needs?

As Christina Soh and Siew Kien Sia point out (MIS Quarterly, 2005), vendors create enterprise systems based around a number of common structures - “ES packages are not custom-built for each implementing organisation”.

“Vendors must make many assumptions about organisational requirements in such areas as organisational policies, structures, standard operating procedures, user knowledge, and interfaces. These assumptions manifest themselves in the processes and features in the ES package”, which the authors refer to as ‘package-embedded structures’.

“ES vendors claim that their package-embedded structures reflect best practice, However, many customers have found that these configuration options do not meet all their specific needs, and many question whether the ‘best practices’ truly do apply to all organisations.

“Developers’ context - that is, their reference organisations - may differ from potential implementers’ contexts, particularly those located in different countries or industries. Even within the same country and industry, contextual differences can exist.”

The system should be able to provide functionality for all of your current and future business processes. To ascertain that this is the case, you first need to define and prioritise your company’s processes, identifying the core business functions and developing a comprehensive requirements list based on input from all stakeholders.

This means that, as Soh et al recommend, “implementing organisations identify, as early as possible, misfits between the package and their organisation. They should create a basis for ascertaining when to align through organisational adaptation and when to align through package customisation.”

‘Misfits’, missing critical features or unsupported business processes, could be the elements that transform an otherwise great fit into a complete mismatch. Very often, these only surface upon implementation.

Buyers should be very wary of future promises from software vendors. If the system does not have the necessary functionality right now in the current release, then you should discount any claims of functionality being available in the future.

The Aberdeen survey warns that, while functionality may be the top selection criterion, “ERP is often considered a commodity today. Don’t assume the functionality you need is available. Take a ’show me’ attitude in demonstration.”

One who has documented this ‘road test’ guide to assess the suitability of a specific solution is Esref Akpinar (2005), who describes a software selection process for a liner shipping company using fuzzy logic decision making. This entailed five scripted scenarios to understand how software packages would handle specific key operational situations. A demonstration evaluation document was prepared, with every question in the document given a weighting according to their importance. An evaluation table was prepared of the results of the demonstrations which clearly indicated which product best fitted the company’s operations and requirements.

2. Total cost of ownership

Prospective buyers should ensure they fully understand the true cost of ownership beyond the initial software licence fees and hardware cost. These may include costs such as those for integration, interfaces, systems communications , extra staff required, upgrades and helpline support. This topic is so important that it has been covered in great detail by the second article in this series, “Managing The Total Cost Of Ownership - What You Need To Know”.

3. Operational Metrics

It is imperative to ensure that not only the costs, but also the benefits of an ERP system are controlled and measured during the implementation project. The benefits generally come in the form of cost savings and operational improvements (e.g. lead time reduction). Cost savings should be built into budgets and operation measures progressively tracked.

Legacy systems often do not support the operational metrics and these have to be assessed manually. A key selection criterion for the new system is thus also the ability to support these operational metrics.

4. Flexibility

Can the application be modified and scaled according to the changing needs of a dynamic and growing business?

Look for an ERP solution that will accommodate new operating protocols, future business growth, market expansion and any other initiatives that might arise.

Things to consider when evaluating flexibility:

• System parameters and default settings;

• Customer screen and menu options;

• Tools for modifying standard forms;

• Data access options and custom reporting; and modular format.

5. Time and ease of implementation

Key questions you should ask regarding the implementation process itself include:

• How long will it take to implement the ERP system?

• Will it cause any major disruptions to your normal business operations?

• Is there an implementation control process (ICP) in place to manage this?

• What sort of business process re-engineering will be required in order to implement the system?

• How long will it take to train staff to use the system?

6. Vendor support and relationships

Your software vendor decision is one that, hopefully, continues well beyond the normal, five-year decision cycle. To that end, three questions are important:

• Does the vendor have a sustainable presence backed up by experience in your industry and a proven track record on installations to similar sized organisations as your own?

• Will you and your management team have a comfortable working relationship that extends to their knowing you and your business intimately? Do they show a sense of responsibility and accountability for making your system choice a success?

• Do you have ‘one throat to choke’? In the event that issues need to be resolved, do you have a direct executive contact who is accountable for making sure your customer service experience is consistently at the highest level?

• How many total vendors will you deal with on your ERP package? Sustaining multiple vendors is cumbersome.

7.I ndustry expertise and customer references

Key questions that will determine the reliability of the vendor include:

• Does the vendor have a proven track record in your specific industry?

• Can the vendor point to a number of companies in your industry who are already using the software and who will confirm that they made a sound decision.

One issue that cuts across many of these selection criteria is the issue of customisation of the software, so it is worth briefly flagging the topic in this context. As Aberdeen Group points out (July 2007), only 11 per cent of respondents to one of its surveys got away with zero customisation.

According to Soh et al, the simplest form of implementation - so-called ‘vanilla’ implementation - requires the organisation to bend to accommodate the software package. “[Vanilla] promotes organisational adaptation, either by conscious redesign and substantial change management, or by piecemeal, evolutionary workarounds, such as individuals and groups adapting. Their adapted practices lead to new organisational structures. Package [software] modification can range from customising the package code to interfacing with custom-developed modules or modules from other vendors.

“Users tend to push for package modification to minimise the amount of change they will have to make. Consultants and project managers tend to advocate organisational adaptation, to simplify package implementation and avoid the tangible costs (time, resources and risks) of package modification.”

The playoff between these two apparently conflicting viewpoints can be a key ingredient to the success of the ERP project, particularly the total cost of ownership, and should therefore be a prime consideration among your selection criteria.

References:

•Akpinar, E., “Software selection for a liner shipping company using fuzzy logic decision making”, paper submitted to the Institute for Graduate Studies in Science & Engineering, Systems and Control Engineering, Bogazici University, 2005

•IBS, “6 Essential considerations when selecting an ERP system”, IBS Australia, February 2008

•Jutras, C., and Dalle Tezze, H., “When relacing ERP - Size matters”, Aberdeen Group, June 2007

•Jutras, C., Trost, J., and Dalle Tezze, H., “Taking the ERP plunge for the first time”, July 2007

•Soh, C., and Siew Kien Sia, “The challenges of implementing ‘vanilla’ versions of enterprise software”, MIS Quarterly Executive, September 2005



By: Peter Clarke

About the Author:

Angela