xi To be able to plan, one needs intelligence. To act intelligently, however, one needs to plan. The questions that this paradox raises express the goal of this book. The abilities to anticipate and plan are essential features of intelligent system, whether they are human or machine. We might go further and contemplate that a better planning results in higher achievements. As a consequence, understanding and improving planning is important.

[JLJ - To be able to plan, one first needs to be able to understand the driving forces which exist and which will therefore naturally hinder any effort to change the present conditions. One also needs a repertoire of schemes of action which one can execute. Such intelligently crafted schemes of action will require to some extent an examination of the likely consequences of possible events, and the necessary rearrangement of resources. Such back-of-the-envelope exercises to diagnostically test one's capacity to adapt and therefore to practically 'go on,' which written out or perhaps viewed from a different point of view, are called plans. Perhaps intelligence arises indirectly from - among other things - the ability to reduce the word of one's predicament to interacting symbols with observable cues and expected interacting effects, which can be anticipated and therefore managed, allowing one to plan effectively...?]

p.1 [Jacob L. Crane] No living thing seems to be conscious of the future, and none seems concerned to design for that future, except Man. But every man looks ahead and attempts to organize for tomorrow, the future of the next day or of the next generation. Whatever he has to do or proposes to do, he plans; he is a planner... By his very nature every man plans constantly.

[JLJ - ...by his very nature and the nature of his predicament, man discovers with maturity that a strategic approach to his problems - coupled with a planned or ordered sequencing of events (and a hefty dose of adaptive capacity) - results more often than not in progress towards desired goals, even if incremental. The exception to this occurs in competition, where even best efforts and good strategic plans may fail to work against those who can outmaneuver at every level.]

p.2 As a starting point, we presume that all intelligent systems use anticipation to plan (van Wezel and Jorna, 2001). An anticipatory system is "a system containing a predictive model of itself and/or its environment, which allows it to change state at an instant in accordance with the model's predictions pertaining to a later instant" (Rosen, 1985).

[JLJ - As a starting point, it would be better to assume that those in a predicament observe the effects of interacting forces over time and often scheme for ways to improve their situation. Some of these schemes will never work while it might occur to us that other schemes - especially those copied from the successful or presented as case studies in business school - are possible, even probable. To formally propose such a scheme as a way to operate, going forward, and to defend it against detractors, is to plan.]

p.2-3 Our definition of planning will be built around this definition of anticipatory system, by distinguishing three main elements of planning.

First, it is important to acknowledge that some entity must make the plan... The planning entity needs some kind of model of the future... Planning and anticipation presume that such a predictive model is available... Information must be collected..

Second, someone or something must execute the plan

The third element of planning is the plan itself... It can never be a full specification of the future itself because it can never be specified more precisely than the model of the future allows.

[JLJ - The plan must be paired with - and can only exist for - the entity in a predicament, with a ticking clock and the presence of competitor(s) and counter-planning-other-entities, and a real world with driving forces that can undermine and even kill you, and which all contribute to the critical need for a plan. A certain amount of adaptive capacity, experience, and convertable resources such as money, will allow the plan to adapt as necessary, and to defer for a later time certain specific details, which will be generated if needed at a later time.]

p.4 Somehow, intelligent systems know how to make planning simple enough to be manageable but complex enough to attain advantageous goals.

p.4 Simon (1981) notes that complex systems are usually somehow ordered hierarchically in order to manage complexity.

p.4 Planning problems are often transcomputational, which means that the amount of plan alternatives is so large that even a computer that is the size of the earth cannot assess all possibilities in millions of years (Klir, 1991).

[JLJ - ...yet there are often heuristics which can convert the planning problem to a simpler model, or even order trial-and-error 'explorations' of the alternatives and interactive effects, so that the solution, while imperfect, is often acceptable, usable, or changeable on-the-fly, and therefore practical. Whatever the planning problem, there is usually a successful planner in that domain who has written a book or perhaps has retired to university to teach, in which case, one simply has to educate oneself in such schemes and techniques, perhaps using the business school model of 'case studies,' and then do what the successful 'did', to 'get' the results they 'got'.]

p.7 First, planning is a synthetic rather than an analytic (diagnosis) or modification (repair) task. Second, planning involves decisions about the future and not the execution of these decisions. Third, an important feature of planning is that it is about choosing one alternative out of a huge number of alternatives that are structurally similar.

p.17 Our starting point in this book is that planning is always in essence about the same thing: anticipating on the future and determining courses of action.

[JLJ - Yes, but on a higher level planning is also in the end about wins and losses - a win when a customer chooses to buy your product over that of the competition or arrives at their destination on time when using your transportation system, a loss when a sports team misses a final field goal attempt, a win when the board of directors decides to continue with the current CEO because they support his grand vision for the busness. We plan in order to win in a competitive event - perhaps not all the time, but often enough to satisfy ourselves that our efforts were worth the time and money spent.]

p.20-22 There are two parts to this book: Part A is theoretical and part B is practical. Part A contains 10 theoretical chapters... Part II of this book contains descriptions of different planning approaches that are applied to one case study, namely the shunting of trains at a station of the Netherlands Railways... In the conclusions (Chapter 18), we assess the various planning approaches that are discussed in this book

p.37 the ability to acquire action-effect associations does indeed seem to precede the capacity to make intentional use of them.

[JLJ - You really cannot form plans of any kind without a working repertoire of action-effect associations.]

p.45 an action plan is actually an integrated assembly of (previously) integrated assemblies.

[JLJ - ...alternatively, one could argue that an action plan is a formal or improvised modification to a previously designed action plan, developed from a position where one has the necessary or available information resources, acquired experience, and skilled capacity to do so...]

p.46 Commonly, we have planned how to react long before the relevant stimulus conditions occur

[JLJ - ...perhaps this is how computer chess programs 'play' chess. We program a very specific, yet complex sensitivity, and then let the machine 'react' to what it 'sees' - the pieces on the gameboard and their present and anticipated complex interactions.]

p.50 Planning is a double-edged sword. On the positive side, it makes creative use of the anticipations, expectations, and future possibilities and thereby strongly enhances the temporal operation space of purposeful action. On the negative side, however, planning an action strongly relies on the accuracy of those predictions of future events and, thus, is likely to fail if some of them turn out to be invalid.

[JLJ - Yes, but we can 'plan' to be practically adaptive, with a useful adaptive capacity (perhaps confirmed via diagnostc tests, such as fire drills or a safety crash rating for vehicles), and make sure that we have multiple valid ways to proceed that appear to be independent of each other. That way if one of them 'fails' due to unforeseen events, another can be selected if it achieves much the same effect or result. Certain 'plans' are doomed to fail to achieve 100% success the minute they are implemented - such as the tax code. People subject to taxation search for clever ways around the 'plan' to collect money in taxes. These 'tax shelters' are then deliberately closed, but new schemes aimed at tax avoidance emerge. 'Plans' meet and clash with other people's 'plans', and there must be a 'survival of the cleverest'. The defense confronting an opposing team in the game of American football has a general 'plan' to physically oppose their opponent, whatever it is that is presented, whatever formation used. A general scheme is modified, then modified again as the offense adjusts. Then as the play develops, everything is improvised based on drills in practice. The 'plan' as such is to develop an experienced scheme of physical opposition that is never caught unprepared, capable of improvising on the fly if necessary as the situation develops, based on cues and general observations made ahead of time.]

p.62 In fact, planning is no longer an end in itself, but rather a means to reach supervision and control goals... a plan does not include all the details needed for its implementation... a plan always integrates a projection in the future

p.63 Reason (1990)... His theory considers that schemata, which are always triggered on the basis of insufficient cues, guide behavior. This underspecification at the same time explains most of the human errors, and the heuristic value of human cognition

p.65-66 At the symbolic level, anticipation is elaborating a representation of the future (forecasting)... expert process operators have more planning ability than beginners because they anticipate more and because they have access to a more global and functional (abstract) representation of the system.

p.67 Bainbridge (1978) has often stressed the difference between experts and beginners when controlling dynamic processes with time lags, on the basis of observations in the field. Experts are able to consider inertia and response delays to produce a restricted number of well-calibrated actions at the right time. On the contrary, beginners act wildly on the basis of immediate but irrelevant feedback, as if the process would be fully controlled by their actions.

p.68 ini dynamic situations... meaningful procedural plans necessarily integrate human actions, goals, intentions, and so on, but also the spontaneous process evolution.

p.69-70 Rasmussen's step ladder has been widely used to describe diagnosis and planning in dynamic situation management... However, Hoc et al. (1995) consider that this model is too procedural to account for every type of dynamic situation.

p.71-72 Our Dynamic Situation Management architecture... there is a shift from determination by a procedure to control by a mental model of the situation. Such a representation is fed by knowledge..., by process information, and by the emergence of subsymbolic entities into the attentional field... It is especially composed of plans defined at various abstraction levels.

[JLJ - I would say that one needs competition-tested schemes which draw our attention to important cues (of all kinds and types) which inform us indirectly what is happening, or to those indeterminate cues which indicate that we simply do not know and likely need to actively gather information to resolve ambiguous cues.]

p.96 Planning includes abstraction - that is, the use of overall representations of situations and actions stressing the crucial properties... The question is finding the right abstraction level to guide the individual or the collective activity reducing the complexity of the data, allowing some degrees of freedom and providing resources for adaptation to circumstances.

p.96-97 Anticipation is costly and limited by time constraints. It can be a handicap when it goes beyond actual competency, so that reactive strategies are necessary. The focus on planning strategies should not let us forget the adaptive power of reactive strategies, especially when time constraints are high or process knowledge is weak. The main problem to solve is finding an efficient compromise between anticipative and reactive strategies to reach a satisfying performance level.

[JLJ - Yet in some fields such as running a nuclear power plant, perhaps one must out of necessity choose plans and strategies where little is left to chance.]

p.97 Adaptation to a situation can be very successful without forecasting it explicitly. A large part of the cognitive activity is regulated at a subsymbolic level without emerging within the attention field.

p.109 In the Encyclopedia of Cognitive Science a problem is defined as "transforming a given situation into a desired situation or goal" (Wilson and Keil, 1999, p.674). Problem solving is the recognition of a discrepancy between an existing and desired situation for which no automated series of actions exist to resolve that discrepancy. Otherwise, it is an algorithm... decision making is the process of choosing a preferred option or course of action from among a set of alternatives"

p.109 we believe that it is not possible to give definitions of planning, problem solving, and decision making that clearly separate one from the other.

[JLJ - Planning as such must be part of a strategic scheme, where available resources are organized and intermediate steps are synthesized. A plan without a practical strategic scheme that calls for the plan, is detatched from reality to a large extent, and therefore is unlikely to produce any effective results - kind of worthless. I can create a plan to be a millionaire in 10 years by acquiring 100,000 dollars each year, but if I do not have a strategic scheme which can net me that amount each year, that I can practically execute, my plan is not - and will not be - effectively realized.]

p.110 The common sense idea is that planning, problem solving, and decision making are overlapping. They are not the same, but they have various activities, starting points, and aspects in common.

p.112 Hierarchical planning means that there is a nested number of goal and subgoal structures or a hierarchy of representations of a plan. The highest level in the hierarchy may be a simplification or an abstraction, whereas the lowest level is a concrete sequence of actions to solve (a part of) the planning problem.

p.113 Planners do not have an in-depth and complete overview, but act according to changing states of affairs.

p.118-119 Mietus (1994) showed that planning as an integrated task consists of various subtasks such as counting, negotiating, evaluating, and problem solving itself. She found that in some cases simple counting takes more than 15% of the time to solve the planning problem... Our expectation... is that planners will use simple schemes without much quantification. This is in line with many remarks from practice saying that planners normally make their plans and schedules on the back of an envelope.

[JLJ - ...provided that such techniques actually work...]

p.139 Organizations that have to function in the real world and in real time have to cope with risk, uncertainty, imperfect knowledge, bounded rationality, and limited communication.

[JLJ - As opposed to organizations that do not have to function in the real world or in real time...? Duh...you have to use what you have, or can borrow, or can invest, in order to go on, wherever you are, and whatever the conditions are or will be. Quit your whining...]

p.141 Coordinated action is characterized by task fulfillment where actions are synchronized to reach augmentative effects... A plan is a social construct.

p.143 Multi-actor systems belong to the physical world as well as to the semiotic world.

p.145 The environment that humans and animals live in can be seen as a physical world and as a semiotic world.

[JLJ - In the human mind, the real is intertwined out of necessity with the practically symbolic, due to the demands of the predicament we live in, and the constant need to synthesize how to 'go on,' from an environment of constant change and the presence of powerfully persistent, often clashing and driving forces. What is real for us is often the symbolic reduction of the world into comfortable semiotic pieces which we can play with, that we rearrange and assemble like a child playing with blocks, in order to understand, and in order to do what is necessary, which is to synthesize how to 'go on' from a world which very often presents us with no obvious directional arrows.]

p.157 Suchman... says that plans are often used as a rationalization afterwards for actions that in reality have been chosen in a decentralized and adhoc manner, based on the situation at hand and some general rules of behavior. She calls this form of action choice situated action.

p.167 Planning is a form of coordination. Coordinated action leads to benefits because of...

1. a better utilization of the capacities of actors...
2. a better utilization of the specialization of actors and means of production...
3. a better balance of the interests of the participating actors...
4. the avoidance of destructive deadlocks and conflicts.

p.177 Planning is a basic cognitive ability that humans use to organize their life and to achieve objectives.

[JLJ - I would say that scheming to achieve goals is a basic cognitive ability, and which most frequently requires planning in order to become effective.]

p.220 algorithms are still based on domain analyses... when creating an algorithm, the developer looks at the characteristics of the "objects" that must be attuned... In general, there are three phases in using such algorithms:

1. Translation of the Domain into the Model...
2. Solving the Model...
3. Translation of the Solution to the Domain

p.224 Planning always somehow deals with arranging scarce resources.

[JLJ - Yes, but consider that a board of directors - in order to run a company - simply has to interview candidates for Chief Executive Officer (CEO) and pick one with a grand vision, experience, and acceptable salary demands. The board of directors then monitors their selection as he/she runs the organization, and replaces the CEO when performance is not acceptible, or stockholders elect a new board with a new vision. You have to either plan, or hire someone who can accomplish the task of planning. Either way.]

p.225 Models of objects at different hierarchical levels are linked to each other by constraints. Constraints restrict the solution space of a decision. In an object model, constrints depict illegal combinations of object instances. Often, there are also rules that express preferences rather than illegal combinations. Such preferences are called goal functions.

p.228-229 As discussed in Chapter 1, planning is a recursive process of setting constraints and goals for lower, more detailed planning levels.

p.236 An algorithm should not be created for a planning problem, but instead for a planner's subtask.

p.240 The position we take is that closing the gap should start with empirical research.

[JLJ - Of course... a PhDs solution to any problem that they do not know the answer to, that interests them, is that we need to study it more. A more practical approach would be to announce a competition of some kind, reward the entrants with money if not enough show up, and see what starting points and approaches you get.]

p.246 a planning process is interpreted as a decision-making process aiming at the selection of appropriate actions that are to be carried out in practice such that the planned system will achieve certain objectives

p.259 Heuristics are solution methods that are usually based on approximation techniques, rules of thumb, or on a decomposition of the problem into less complex subproblems.

[JLJ - Heuristics might be useful to construct a first-order solution to a problem, which can then either guide actions now, or in other situations, can be refined later if a more accurate solution is needed.]

p.275 the mathematical models provide the planners with useful initial solutions, but these initial solutions have to be fine-tuned and fit together manually in a second step.

p.276 Mathematical models and solution techniques can be used to generate solutions or partial solutions for complex planning problems. These solutions may be used directly in practice, or the planners may further refine them manually if they consider this necessary.

p.276 developing a mathematical model requires one to balance the required level of detail of the model carefully with the implied complexity. However, in practice there is often a tendency to incorporate as many details as possible into a model. The latter may have a detrimental effect on the model's complexity. Creating mathematical models and corresponding solution techniques that are really useful in practice is therefore more of an art than a science.

[JLJ - ..that is, unless one has a mechanism to automatically test the effecitveness of a model. One can then change the model and observe the effects on performance, either intelligently or randomly. This is hardly an 'art.']

p.303 The effect of a nondeterministic action is given as a set of possible next states.

p.304 An adversarial planning domain has a single system agent that is controllable and a single environmental agent that is uncontrollable and may be an opponent to the system agent. The task is to synthesize plans for the system agent that is robust to any plan of the environmental agent.

p.331 Absence of "planning" in most cases means no intelligence.

[JLJ - It could very well be that the absence of planning is acceptible, as in a 7 year old playing with blocks has no need or desire to plan, because it is acceptible for his or her parents to do planning for him. Planning becomes a necessary formality for organizations that wish to exist in the future, and for those who want a better life than what is generally available in now taking a job here, now taking a job there, with no formal schooling. Serious time spend planning is a strategic and intelligent approach to how to 'go on,' when one wants better than average results to arrive, when the future becomes the present. Plan all you want - for as long as you want - if you do not understand the driving forces of your predicament, if you cannot manage these driving forces along with satisifying the needs of those in your organization, and work with outside organizations, your plans, whatever they are, will likely fail.]

p.334 We need to reduce the complexity of computations by grouping similar units (entities) into a larger formation that can satisfy the definition of an entity, too.

[JLJ - I would use the word 'ought', or insert the word 'perhaps'.]

p.336 The abbreviation GFACS is deciphered as "Grouping, Focusing Attention, and Combinatorial Search"... Most of the elementary procedures that are being applied for... other intelligent activities are, in turn, based on the GFACS set of procedures.

p.340 Intelligence is a property of the system that emerges when the procedures of direct and inverse generalization (including focusing attention, combinatorial search, and grouping) transform the available information in order to produce the process of successful system functioning... the main advantage of intelligence is the ability to deal with unexpected predicaments.

[JLJ - ...we are always in a predicament, with much uncertainty. Perhaps 'the main advantage of intelligence' is also the ability to manage the predicament, which includes taking advantage of unexpected opportunities.]

p.344 semiotics is a science of signs. Why signs? The signs transmit compressed information, or messages; the messages contain meanings concerning the events and processes; the meanings are constructed and produced by intelligence from signs and the inner knowledge... signs and their processing (the body of semiotics) embody intelligence (the machinery for simulation).

p.348 Each scale has its system of signs and is related to the adjacent scales by the set of rules that represent relationships of generalization (bottom-up) and instantiation (top-down) that exist between the scales.

p.352 Cultivating the accepted degree of redundancy is a prerequisite of intelligence. Redundancy of systems is understood as having their resources, components, or properties in abundance, or in excess. This property of redundancy is very important and very characteristic for intelligent systems. They should always be ready to withstand uncertainty, and since the survival is at stake, the property of redundancy helps to minimize the risk of failure.

[JLJ - Yes things somehow should be redundant, but ideally as part of an intelligent scheme to achieve adaptive capacity, so we can improvise as necessary, in order to 'go on.' Perhaps in cases where a fragile scheme cannot yet be exploited, we do not (yet) need to have redundancy. But in the general case, we move forward into the unknown, within our predicament, reassuring ourselves, that whatever it is that awaits us, we are or will be 'ready' for it.]

p.353 only intelligent systems do not fight redundancy, but rather explore, use, and even cultivate the redundancy. Redundancy is the tool for combining and testing new alternatives of decisions and, eventually, novel behaviors... How is it possible to cultivate redundancy and yet fight complexity? ...the tools of complexity reduction should not curb the combinatorial capabilities of the system.

p.353 [footnote] "The system is intelligent if it has more intelligence than it needs."

p.354 In all systems... the formulation of multiresolutional representation is a technique of complexity reduction.

p.365 Part II Practical

[p.365-496 boring study, yawn. Authors of texts forget that their readers might not be students in a classroom that they can force to listen to them pratter about how great they are and what wonderful problems they have solved. Read this section if you have trouble falling asleep.]

p.497 In this concluding chapter we try to structure and (partly) integrate the various perspectives and approaches regarding planning that have been described in the preceding chapters.

p.503 Newell and Simon assume the following steps in planning: (1) abstracting by omitting certain details of the original objects and operators, (2) forming the corresponding problem in the abstract problem space, (3) when the abstract problem has been solved, using its solution to provide a plan for solving the original problem, (4) translating the plan back into the original problem space and executing it

[JLJ - Yes, but it takes intelligence to do all of this, or at minimum, an intelligently designed scheme we can execute. Worse, we must take into account the actions of other intelligent agents, and unknown driving forces that clash, and what emerges from all this conflict condenses, shapes and defines our society and age, and we must act and maneuver within it.]

p.504 In his [Sacerdoti's] view, planning is performed by recursively decomposing goals into subgoals, until a subgoal can be reached by elementary actions.

p.505 In this model, planning is about how to find the actions that solve a problem, or more generally, reach a goal... The process itself is about formulating goals, finding similar solved goals, finding existing plans, adapting plans, learning, and storing plans in such a way that they can easily be found for future use.

[JLJ - Yet there are indirect ways to indefinite goals. As a teenager one might have no idea what to do for a living. An effective plan would be to attend in school and to identify the subjects that you received the best grades in (as well as trying to improve in areas you did not do so well). You would then examine what professions use those skills, and perhaps what univerities offer degrees in those areas. With a ticking clock of time and much unknowable, we often cannot form specific plans, but instead must plan in general, or plan to make the most effective use of time available, then deal with what emerges in the future. I doubt there is a way to 'solve' the problem of what to do for a living, other than to make incremental progress, as time and effort, and opportunities allow.]

p.505 To make a plan, an actor somehow anticipates the future by simulating the actions he will make. The plan is virtually executed in the model of the future... This requires the existence of (internal) representations.. In this paradigm, planning is searching for a sequence of actions that will bring the agent from its current state to the goal state.

[JLJ - Yet you cannot simulate such actions if you have no ideas for what actions are possible. Only after an apprenticeship of some kind is one able to understand enough of the predicament to simulate the actions in the future, in that profession, and to visualize one's place in it.]

p.512 Planning requires intelligence and learning. It requires that entities that plan are in a sense rational, have representations, process information and can anticipate.

[JLJ - Effective planning requires intelligence and learning, in a complex or competitive world.]

p.517 Meystel... states that intelligent systems... are capable of planning because they are able to simulate the effects of decisions before the decisions are actually implemented. Here, "to simulate" means to anticipate and to model... Planning is understood as searching for appropriate future actions leading to the goal... Searching is perfrmed within the system of representation.

[JLJ - Is it searching or is it the intelligent and effective construction of a diagnostic test in order to estimate the ability to adapt, in the unknown future?]

p.522 The larger the space of search, the higher the complexity of search. This is why a special effort is allocated with and focused upon reducing the space of search. This effort is called focusing attention

p.527 planning is inseparable from and complementary to learning

p.527-528 One of the ways to make sure that the actions will have the desired effects is to simulate actions before they are performed - that is, imagining what effects the actions will have. Here, "simulation" means the consideration of possible scenarios without actually performing or executing them. We deliberate and devise. This is the core of planning. We simulate actions, compare the outcome of the simulation with the desired outcome, revise the planned actions, simulate and compare again, and so on, until finally the planned actions will be performed.

p.529 The main task of an intelligent system is to intelligently determine what options will be looked at... the building blocks for this are grouping, focusing attention, and [JLJ - combinatorial] searching (GFACS) on multiple levels of resolution.

p.529 Planning for yourself is considered to be problem solving... planning is defined as search for a trajectory through space... looking for a solution in a chess game may also be conceptualized as searching through a problem or conceptual space.