2/1/2006 | 3 MINUTE READ

Solving Problem Solving

Facebook Share Icon LinkedIn Share Icon Twitter Share Icon Share by EMail icon Print Icon

This problem-solving tool scans the world (ok, the world wide web) for possible solutions to not only bring new ideas to the table, but to help solve problems faster.


Facebook Share Icon LinkedIn Share Icon Twitter Share Icon Share by EMail icon Print Icon

Problem solving has (at least) three problems:
1. It requires information—information that’s made of syntax, semantics, context, relevance, fragments and sometimes fully formed concepts.
2. It’s often a messy process.
3. Shortcuts often result in merely reinventing the wheel.

Enter Goldfire Innovator from Invention Machine Corp. (Boston, MA; www.invention-machine.com), software that brings structure and repeatability to problem solving. It also retrieves answers—prior knowledge, really—to questions hatched in the problem-solving process. Basically, it recasts a problem to be solved in conceptual abstract language, and then finds solutions in corporate and worldwide technical literature. In so doing, says Stephen Brown, Invention Machine’s vice president of strategic marketing, “it improves the odds of converting an idea into a product.” 

Goldfire is best at solving non-deterministic problems, those that don’t have an explicit road map or formula leading to a correct answer. “They tend to be in this space of subjectivity, what-if, and what-am-I-going-to-do-now kinds of head scratching,” explains Brown. Non-deterministic problems are typically solved through brainstorming: people gather together, throw out ideas, and hope something comes out. Naturally, there are problems with this approach. First, brainstorming often depends on the personalities involved, their expertise, and such not-so-extraneous factors as office politics. Second, says Brown, “What you know is limited to what you know.” As a workforce gets older, the more its what-you-know knowledge retires from the enterprise. Last, brainstorming is undirected, unstructured, and not repeatable. Executives are “very uncomfortable at just throwing money into research,” Brown continues; “they don’t necessarily know what they’re going to get out the other end.”



Research as far back as the 1940s in the United States, Russia, and elsewhere has shown that tough technical constraints and tradeoffs are often solved in similar ways. From that has come various formal and provable methods for problem solving. Of note are value engineering, root cause analysis, failure mode effects analysis, and inventive problem solving (TRIZ). These methods, all incorporated in Goldfire Innovator, both guide analysis and stimulate innovative thinking along structured, disciplined ways. The software also incorporates an assumption: There’s nothing new under the sun. Breakthroughs often happen when known science from a different scientific discipline is applied to the problem at hand. In other words, explains Brown, “Real innovation happens not because we create new science, but because we apply cross-disciplinary science in new ways.”

In addition, Goldfire Innovator tackles the challenge of locating desired concepts in a precise and timely manner. Locating known science is supported with worldwide concept retrieval. For starters, Goldfire Innovator’s semantic engine retrieves relevant prior knowledge from structured and unstructured corporate content on shared drives and databases. That in itself is no easy feat within documents across project or departmental silos. The software also retrieves external semantically relevant content on the Web. This process has its own difficulties, as anyone using Google knows. Last, Invention Machine provides semantic access to over 15-million worldwide patents—a treasure trove of ideas, scientific effects, and background “on how systems and components interact and behave functionally to achieve certain outcomes,” says Brown.

Goldfire Innovator’s “semantic knowledge engine” retrieves information relevant to the problem at hand and indexes the content. This is more than a keyword search. Invention Machines linguistic technology “understands” the functional meaning of each sentence in a document. These understandings are then presented to users. Contrast this with traditional search technologies, which neither understand the structure nor the grammar of language. Text string matches miss syntax, context, and concepts that are out of context. Relevant and casual information is missed.



The software costs about $75,000 for 10 seats. Payback can be realized not only in solving problems related to, say, warranty, but also as regards the speed that engineers can get the job done. For example, Delphi had a stress fatigue problem that caused a release actuator for an integrated cinching latch to break after 18,000 cycles. Using traditional root-cause methods, Delphi investigated and resolved the problem in a month. Another Delphi team used Goldfire Innovator. First, they identified all the components relevant to the actuator’s capabilities. Goldfire Innovator’s Device Improvement Wizard then guided the team in determining the functional interaction of the components. The software applied value engineering metrics to focus the team on key areas of high-cost, problematic component behavior, technical contradictions, and functional constraints. Design simplification algorithms guided the Delphi engineers through the process of trimming problematic components and producing multiple alternative designs. Last, the software’s Library of Inventive Principles presented conceptual ideas about how analogous technical problems were solved in the past. The team identified the root cause in 10 hours, which could have reduced iterations and the implementation of the solution by three weeks.

Thanks for considering a subscription to Automotive Design & Production. We’re sorry to see you go, but if you change your mind, we’d still love to have you as a reader. Just click here.