Desalination mega-plants are a collection of rather sophisticated engineering systems. No company may build such a mega-plant alone. Hence, subcontracting in desalination mega-projects is a norm and it adds substantially to the project risks.

Ideally, the subcontractor shall plug in/out of the company business painlessly. In other words, a subcontractor should be considered a loosely coupled part of the business.

Screening out of such parts is centered on the following points.

  1. business financial health
  2. business maturity
  3. business performance
  4. the business fit of the project standards, size, and timeline

Financial health evaluation is an attempt to transform into a go/no-go answer the collection of annual data from the balance sheet (total assets, total liabilities, net worth, current assets, current liabilities) and income statement (total revenues, profit before taxes, profits after taxes, total/operating income, etc ). This task is a perfect candidate for machine-learning methods like decision trees or logistic regression.

Business maturity is a broad category covering qualitative data - staff expertise, automation level, management standards, digitization level, and subcontractors net. There is no clear way to turn this data into maturity indicators.

Business performance is easily scored based on quantitative data - non-performing contracts' total size and quantity, pending litigation, total assets, total liabilities, average annual construction turnover, etc.

Business fit to the project configuration is the hardest part of contractor evaluation as it is based on extrapolation of the contractor's past performance ( described in terms of similarity, complexity, methods, and technology ) on the future state.

Contractor management should be a perpetual process, the earlier it starts the more reliable the contractor evaluation is. Such a strategy is a big time saver during the project's execution.

Contractor management is a two-way road. To start subcontracting the company business shall build a web communication channel providing the following online services.

  1. Access to relevant project data
  2. Bidding and contract management
  3. Works and resources scheduling
  4. Monitoring of subcontracted work progress
  5. Documentation management
  6. Change management
This group of tasks is relatively simple compared to the first group, which leads us to the artificial intelligence domain.

Undoubtedly, its central part is NLP - natural language processing as most of the information about contractors is available as text.

It may describe personnel, equipment, and facilities, outline management roles and responsibilities, and manufacture methods and procedures. The Internet is another huge source of textual information.

Texts may be analyzed to rank the business competition strength, financial performance, vision, strategy, sustainability, etc.

Works similarity assessment is an example of the NLP application. Its core is the text classification problem. The NLP algorithm analyzes the descriptions of the work scopes executed by a subcontractor in the past (as a main contractor, partner, or subcontractor) and selects the project work packages matching the contractor's capabilities with the highest probability.

The biggest problem with building such systems in an easy and fast way is a lack of semantic frames or topic models. They are vocabularies describing domain knowledge and a framework for dense vector semantics. The remarkable fact is that business process modeling implemented by generates semantic frames automatically. They are an inherent part of the object model in the object-oriented architecture adopted by

Another enticing area of the NLP application is the extraction of eye-catching outlier paragraphs from legal contracts, agreements, conditions, policies, and regulations or detecting omissions and errors.

© 2024