Skip to content

Latest commit

 

History

History
160 lines (109 loc) · 6.51 KB

File metadata and controls

160 lines (109 loc) · 6.51 KB

Goldratt's Theory of Constraints — Background

The Man

Eliyahu Moshe Goldratt (1947–2011) was an Israeli physicist who became one of the most influential management thinkers of the 20th century. He held a BSc in Physics from Tel Aviv University and MSc and PhD from Bar-Ilan University.

Goldratt's genius was applying scientific thinking — cause and effect, hypothesis testing, and systematic reasoning — to business problems that most people treated as "soft" or "intuitive."

The Core Idea

Every system, no matter how complex, has at most a very small number of constraints that limit its output. Focusing improvement efforts on anything OTHER than the constraint is wasted effort.

This is simultaneously obvious and revolutionary. Most organizations spread improvement efforts evenly across all departments, or focus on whoever complains loudest. TOC says: find the ONE thing that limits the whole system, and focus there.

The Books

The Goal (1984)

Goldratt's masterpiece. Written as a novel about Alex Rogo, a plant manager whose factory is about to be shut down. Through conversations with a physicist named Jonah (Goldratt's alter ego), Alex discovers:

  • The Goal of a business is to make money (not "produce products" or "keep people busy")
  • Throughput (rate of making money) matters more than efficiency
  • The bottleneck determines system throughput
  • The Five Focusing Steps for continuous improvement

Key concepts introduced: Throughput, Inventory, Operating Expense, Bottleneck, Drum-Buffer-Rope.

It's Not Luck (1994)

Sequel to The Goal. Three managers face crises and use Thinking Processes to save their divisions:

  • Current Reality Tree (CRT) — "What to change?"
  • Evaporating Cloud (EC) — "What to change to?"
  • Future Reality Tree (FRT) — "How to validate the change?"
  • Prerequisite Tree (PRT) — "How to overcome obstacles?"
  • Transition Tree (TT) — "How to implement step by step?"

Also introduces the Layers of Resistance framework for change management.

Critical Chain (1997)

Applies TOC to project management. Key insights:

  • Individual task safety margins are WASTED (Student Syndrome, Parkinson's Law)
  • Critical Chain = Critical Path + resource conflicts
  • Pool safety into project buffers instead of padding tasks
  • Manage by buffer consumption, not by task due dates
  • Stagger projects to avoid multitasking

The Haystack Syndrome (1990)

Addresses information and measurement. Key insight:

  • Throughput Accounting replaces cost accounting for decisions
  • Three measures: T (Throughput), I (Investment), OE (Operating Expense)
  • Product profitability should be measured by T, not by allocated cost
  • T/CU (Throughput per Constraint Unit) determines product priority

Necessary But Not Sufficient (2000)

About ERP and technology implementation. Key insight:

  • Technology is necessary but not sufficient for competitive advantage
  • Without changing rules (policies, measurements), technology just automates waste
  • Three questions: What is the power of the technology? What limitation does it diminish? What rules existed because of the limitation?

The Choice (2008)

Goldratt's most philosophical book. Two thinking tools:

  1. Never say "I know" — always question assumptions
  2. Inherent simplicity — every complex system is governed by very few elements
  3. Every conflict can be resolved — if you dig deep enough into assumptions
  4. People are not stupid — resistance always has a logical reason

Isn't It Obvious? (2009)

Applies TOC to retail and distribution:

  • Replenishment instead of forecast-based ordering
  • Daily replenishment from regional warehouses
  • Hold inventory at the aggregate level (warehouse), not at each store
  • Measure by Throughput Dollar Days (TDD) and Inventory Dollar Days (IDD)

The Thinking Process Tools

The six tools form a complete problem-solving methodology:

"What to change?"
  └── Current Reality Tree (CRT)
       Maps cause-effect chains from symptoms to root causes

"What to change to?"  
  ├── Evaporating Cloud (EC)
  │    Resolves the core conflict by breaking hidden assumptions
  │    → Produces an "injection" (the change)
  │
  └── Future Reality Tree (FRT)
       Validates the injection: will it work? What could go wrong?
       → Identifies and trims Negative Branch Reservations (NBRs)

"How to cause the change?"
  ├── Prerequisite Tree (PRT)
  │    Lists obstacles and intermediate objectives
  │
  └── Transition Tree (TT)
       Step-by-step action plan: reality → action → effect → next reality

Logic Foundations

Sufficiency Logic (If...Then...)

Used in CRT, FRT, and TT.

"If [cause], then [effect]" — the cause is sufficient to produce the effect.

Validated by the Categories of Legitimate Reservation (CLR):

  1. Clarity
  2. Entity existence
  3. Causality existence
  4. Cause insufficiency
  5. Additional cause
  6. Cause-effect reversal
  7. Predicted effect existence
  8. Tautology

Necessity Logic (In order to...we must...)

Used in EC and PRT.

"In order to [objective], we must [requirement]" — the requirement is necessary.

Validated by: "Can we achieve [objective] WITHOUT [requirement]?"

Impact and Legacy

TOC has been applied in:

  • Manufacturing: Thousands of factories worldwide
  • Project management: CCPM adopted by military, construction, software
  • Healthcare: Hospital throughput improvement
  • Education: Thinking Processes taught to children
  • Software: Kanban systems draw from TOC principles
  • Supply chain: Replenishment-based distribution

The TOC International Certification Organization (TOCICO) maintains standards and certification.

Why AI + TOC?

Goldratt's tools are powerful but require significant training and practice. The Thinking Processes in particular demand rigorous logical thinking that humans often shortcut.

AI agents are well-suited to TOC because:

  1. Systematic logic: LLMs can rigorously apply If→Then and Must→To chains
  2. Assumption surfacing: AI excels at generating and questioning assumptions
  3. Pattern recognition: Identifying root cause patterns across domains
  4. Patience: Building a full CRT with 15 entities takes discipline AI naturally has
  5. Multi-perspective: Agent teams can represent different stakeholder views

This project (toc-agents) is the first attempt to make TOC's full toolkit accessible through AI, lowering the barrier from "years of training" to "type a command."